Microsoft’s chief scientist has warned that Donald Trump’s proposed ban on state-level guardrails on artificial intelligence will slow the development of the frontier technology rather than accelerate it.

Dr Eric Horvitz, a former technology adviser to Joe Biden, said bans on regulation will “hold us back” and “could be at odds with making good progress on not just advancing the science, but in translating it into practice”.

The Trump administration has proposed a 10-year ban on US states creating “any law or regulation limiting, restricting, or otherwise regulating artificial intelligence models, artificial intelligence systems, or automated decision systems”.

It is driven in part by White House fears China could otherwise win the race to human-level AI, but also pressure from tech investors, such as Andreessen Horowitz, an early investor in Facebook, which argues consumer uses should be regulated rather than research efforts. Its co-founder, the Trump donor Marc Andreessen, said earlier this month that the US was in a two horse race for AI supremacy with China. The US vice-president, JD Vance, recently said: “If we take a pause, does [China] not take a pause? Then we find ourselves … enslaved to [China]-mediated AI.”

Horvitz said he was already concerned about “AI being leveraged for misinformation and inappropriate persuasion” and for its use “for malevolent activities, for example, in the biology biological hazard space”.

Horvitz’s pro-regulation comments came despite reports that Microsoft is part of a Silicon Valley lobbying push with Google, Meta and Amazon, to support the ban on individual US states regulating AI for the next decade which is included in Trump’s budget bill which is passing through Congress.

Microsoft is part of a lobbying drive to urge the US Senate to enact a decade-long moratorium on individual states introducing their own efforts to legislate, the Financial Times reported last week. The ban has been written into Trump’s “big beautiful bill” that he wants passed by Independence Day on 4 July.

Horvitz was speaking at a meeting of the the Association for the Advancement of Artificial Intelligence on Monday when he said: “It’s up to us as scientists to communicate to government agencies, especially those right now who might be making statements about no regulation, [that] this is going to hold us back.

“Guidance, regulation … reliability controls are part of advancing the field, making the field go faster in many ways.”

Speaking at the same seminar, Stuart Russell, the professor of computer science at the University of California, Berkeley, said: “Why would we deliberately allow the release of a technology which even its creators say has a 10% to 30% chance … of causing human extinction? We would never accept anything close to that level of risk for any other technology.”

skip past newsletter promotion Sign up to Business Today Free daily newsletter Get set for the working day – we'll point you to all the business news and analysis you need every morning Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The apparent contradiction between Microsoft’s chief scientist and reports of the company’s lobbying effort comes amid rising fears that unregulated AI development could pose catastrophic risks to humanity and is being driven by companies prioritising short-term profit.

Microsoft has invested $14bn (£10bn) in OpenAI, the developer of ChatGPT, whose chief executive Sam Altman who this week predicted that: “In five or 10 years we will have great human robots and they will just walk down the street doing stuff … I think that would be one of the moments that … will feel the strangest.”

Predictions of when human-level artificial general intelligence (AGI) will be reached vary from a couple of years to decades. The Meta chief scientist, Yann LeCun, has said AGI could be decades away, while last week his boss, Mark Zuckerberg, announced a $15bn investment in a bid to achieve “superintelligence”.

Microsoft declined to comment.

Eric Horvitz’s comments come despite reports that Microsoft is lobbying with Google, Meta and Amazon to support ban

Trump’s plan to ban US states from AI regulation will ‘hold us back’, says Microsoft science chief

Austin, Texas is set to be the first city worldwide to see Tesla’s self-driving robotaxi service on its roads. Elon Musk, CEO of the electric carmaker, has said he is “tentatively” planning to roll out a small number of these autonomous vehicles on the streets of the Texas state capital on Sunday.

Details about the company’s robotaxi service have been scant since its unveiling in October of last year, and its launch has been delayed. Musk has told reporters that there may be fewer than a dozen cars in Austin on Sunday and that the vehicles will stick to specific neighborhoods. Some analysts believe that the robotaxis will only be available to employees and invitees initially.

For the CEO, Tesla’s rollout is slow. “We could start with 1,000 or 10,000 [robotaxis] on day one, but I don’t think that would be prudent,” he told CNBC in May. “So, we will start with probably 10 for a week, then increase it to 20, 30, 40.”

The billionaire has said the driverless cars will be monitored remotely, but it is unclear if a safety driver will accompany the vehicles – as is standard practice when other robotaxi companies have launched in new cities.

Musk has been touting the Austin rollout since last summer, but the debut date has been a moving target. Over the past few months, the billionaire indicated the launch would be some time in June.

Then, news reports floated that it would happen on 12 June. Just two days before that supposed launch, Musk reposted a video on X, the social media platform he owns, showing a driverless Tesla passing through an intersection in Austin. In response to questions about the video on X, Musk said the date was “tentatively” 22 June but that this launch date would be “not real self-driving”, which would have to wait nearly another week.

“We are being super paranoid about safety, so the date could shift,” Musk added. “First Tesla that drives itself from factory end of line all the way to a customer house is 28 June.”

Musk said he plans to have one thousand Tesla robotaxis on Austin roads “within a few months” and then he’ll expand to other cities in Texas and California.

Despite a launch that Musk has described as scaled back, Tesla is already facing opposition from Texas lawmakers. A group of Democratic state senators and members of the house sent Tesla a letter last week asking the company to delay the debut until a new autonomous vehicle law takes effect in September.

“We believe this is in the best interest of both public safety and building public trust in Tesla’s operations,” the lawmakers wrote. It is unlikely their objections will throw up lasting roadblocks, though, as Republicans hold majorities in both chambers of the Texas legislature.

Federal authorities have likewise taken note of the robotaxi launch. In May, the National Highway Traffic Safety Administration sent Tesla a list of questions about its robotaxi launch in Austin, which included how the company would intervene if something went wrong, how it would handle crash reporting and how it would deal with adverse weather conditions.

An NHTSA spokesperson told the Guardian that the agency has received a response from Tesla and it is in the process of reviewing it.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Tesla has been the subject of numerous investigations and lawsuits over the safety of the software that allows owners of its vehicles to drive hands-free, which the company calls Full Self-Driving and which is closely related to the technology used in its robotaxis.

There have been at least 17 deaths and five serious injuries involving the technology, according to the Washington Post. The NHTSA is looking into numerous crashes in which the agency says Full Self-Driving software was impeded by weather conditions like sun glare, fog, dust and darkness, including one crash that resulted in the death of a pedestrian.

Tesla will be joining a crowded robotaxi market in the Texas city, which already has services from Waymo, owned by Google parent Alphabet.

Amazon and Volkswagen are also testing driverless taxis in Austin. These companies have taken a slower approach in developing their autonomous vehicle technology by driving hundreds of thousands of miles to map city streets and working to train their software to avoid pedestrians, cyclists and unexpected happenings on the road.

Tesla did not return a request for comment.

Elon Musk’s autonomous robotaxis slated to hit the streets of the Texas capital on Sunday after delayed launch

Tesla set to unveil self-driving car service in Austin

Thousands of UK government laptops, phones and tablets worth more than £1m have been either lost or stolen, freedom of information disclosures have revealed, triggering warnings of a “systemic risk” to the nation’s cybersecurity.

The Department for Work and Pensions recorded 240 missing laptops and 125 missing phones in 2024; while in the first five months of this year the Ministry of Defence recorded 103 missing laptops and 387 missing phones. The Cabinet Office, which coordinates government activity, lost or had stolen 66 laptops and 124 phones in 2024.

The replacement cost of the more than 2,000 missing devices recorded across 18 Whitehall departments and public authorities in the last year for which figures are available is running at about £1.3m annually, according to Guardian analysis of freedom of information responses. The Bank of England, HM Treasury and the Home Office were among other departments where dozens of phones and laptops went missing.

Cybersecurity experts said the losses could enable hackers to create backdoors into government systems even if large parts of the hardware were encrypted. One called it “a huge national security risk”, but the government downplayed the danger, saying that encryption prevented access to bad actors.

“These are surprisingly large numbers,” said Prof Alan Woodward, a cybersecurity expert at the University of Surrey. “When you are talking about so many [it creates] a large attack surface [for hackers]. If 1% were system administrators who had their phones stolen, that’s enough to get in.”

He said that if devices were open when stolen, as frequently happens with phones snatched on the street, criminals could keep them open and “drill down into the device and once the phone is open, by design it is readable and accessible”.

The Ministry of Defence said it had robust policies and procedures to prevent losses and thefts. It said: “Encryption on devices ensures any data is safeguarded and prevents access to the defence network.”

The Bank of England said it “takes the security of devices and data very seriously and has suitable protection in place”.

A government spokesperson said: “We take the security of government devices extremely seriously, which is why items such as laptops and mobile phones are always encrypted so any loss does not compromise security.”

It added that every loss or theft was investigated.

“The device loss seems quite high,” said Nick Jackson, the chief information security officer at Bitdefender, a cybersecurity firm. “It only takes one lost [device] to compromise a network. It poses a systemic risk and is something that could potentially be taken more seriously especially given the access and connections that department will have.”

skip past newsletter promotion Sign up to First Edition Free daily newsletter Our morning email breaks down the key stories of the day, telling you what’s happening and why it matters Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

He said laptops were likely to have encryption, but tablets or phones presented a greater risk.

Jackson said: “The biggest risk is that the devices themselves will have access to sensitive information and authentication tokens. If someone was able to gain access to those they would be able complete authentication processes on any government application or government website that they shouldn’t be able to access.”

The Department for Science, Innovation and Technology, which is responsible for cybersecurity, recorded 83 phones and 18 laptops lost or stolen in the year to May 2025. In 2024, the Home Office, which oversees policing, had 147 devices go missing at an estimated replacement cost of more than £85,000.

An MoD spokesperson said: “We treat all breaches of security very seriously and we require all suspected breaches to be reported. All incidents are subjected to an initial security risk assessment, with further action taken on a proportionate basis.”

David Gee, the chief marketing officer of Cellebrite, a digital forensics and cybersecurity firm that works with the Metropolitan police, said: “Missing devices pose a huge national security risk, especially coming from public sector departments where they hold vast amounts of sensitive data. From healthcare departments to defence, staff phones and laptops must be protected at all costs, and keeping data safe in these government agencies should be a top priority.”

Exclusive: ‘Systemic risk’ to cybersecurity after devices worth more than £1m a year are taken or lost

Thousands of UK government laptops, phones and tablets have been lost or stolen

In April of this year, Daisy Yelichek was scrolling TikTok when something unusual appeared in her feed: a video of her 84-year-old father, George Tsaftarides, who runs an account sharing sewing videos from his small tailoring business in Ohio. But the video Yelichek was seeing was not from Tsaftarides’ actual page, which has nearly 41,000 followers – but instead originated from a profile of someone claiming to be a “sad old man” whose cat sanctuary was at risk of shutting down.

“Please stay 8 seconds so I don’t have to shut down my cat shelter I poured my love into,” the text on the video said, adding that the sanctuary would be selling slippers to raise additional funds. The bid for sympathy worked on many viewers, garnering millions of views and tens of thousands of users leaving concerned comments. “Just ordered two! Sending love to these kittens,” wrote one. Another commenter said: “thank you for all you do for these babies.” Others even asked if there was a GoFundMe link to donate directly to the cat shelter.

Yelichek and her father were shocked. Tsaftarides does not run a shelter. The account posting the plea for funds appears to be a front for a scheme seeking to sell mass-produced slippers. Several of Tsaftarides’ followers who actually ordered slippers complained that the fuzzy footwear came with “made in China” tags and did not, in fact, appear to be handmade by an elderly man with a struggling cat sanctuary.

“These people are using my identity to make money and I don’t understand why,” Tsaftarides said. “It’s thieving, it’s stealing, and it’s not right.”

Tsaftarides is not the only TikTok user who has had his likeness hijacked by such accounts. Charles Ray, an 85-year-old retiree based in Michigan, has also been targeted by accounts using doctored videos of his likeness. He started his actual TikTok account in January and uses it to share jokes with his followers.

“Earlier this year, my pastor told me a joke about a frog, and I thought, ‘that ought to make people smile,’ so I figured out how to make an account and told the joke, and it took off from there,” he said.

Ray’s videos all follow a similar template: filmed in selfie mode, he tells a short joke. He was frustrated to learn his content, which he makes only to “share joy” and not to earn money, was being lifted and edited to scam people. In one video, Ray is rubbing his eye, and the repost seems to imply he is crying. Another video uses a clip from a woman crying on TikTok about an unrelated issue, and another includes a user in a hospital bed.

Since she first discovered the proliferation and manipulation of her father’s face, Yelichek has identified more than 100 accounts splicing his videos with other, unrelated users to sell mass-produced goods including slippers, headphone cases and blankets – all under the guise of independent sellers that need help. Some two dozen Instagram accounts and YouTube as well as a handful were pulling the same trick, according to a list compiled by Yelichek and a review of the accounts by the Guardian.

At one point, Yelichek even made contact with the account manipulating her father’s likeness over direct message and pleaded with its owner to stop. The person behind the account claimed to be a poor 17-year-old boy based in Greece trying to make money for his family.

“I totally understand your situation but I also want you to understand mine,” he said, proceeding to post more stolen videos.

The con replicates a recurring genre of video on TikTok that has boosted sales for some small businesses: user makes a heartfelt post about a local store or restaurant that is struggling, and online followers are moved to support it. One typical post mimicking a local plea for help reads: “Please just stay 15 seconds to save my pawpaw’s slipper small business.”

In the case of videos Yelichek is seeing, many commenters who are moved by the fake story try to boost it by commenting names of celebrities and creating other engagement they believe TikTok prioritizes. “Fun fact,” a comment on one video of the type reads, this one using the same formula but claiming to be a struggling cow sanctuary. “Liking and replying to comments boosts more! Referencing popular things like Chappell Roan and Taylor Swift puts this video on the [for you page] of more people.” The video pulled in 1.4 million likes and 26,000 comments before being taken down. Yelichek says these accounts lift videos from other TikTok users as well and recontextualize them to create a false narrative.

Tsaftarides said his content being used to sell mass-produced goods is particularly frustrating, as he started his account to promote small businesses, including his own, and to encourage people to shop locally.

“All we want to do is show people our store and teach them about sewing,” he said. “We don’t make money off of our TikTok account.”

Yelichek said she and her father filed a police report to Jackson township police in Ohio, where their store is based, for identity theft and have made great efforts to get social networks to take the stolen content down – often to no avail. Yelichek said that while Instagram has removed a few of the profiles she reported, TikTok – the platform where the issue is more widespread – has been less responsive.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

“If we comment on [these videos] saying they’re spam, our comments often get deleted right away, with TikTok saying it’s against their community guidelines,” Yelichek said, sharing screenshots of the messages. “They’ve actually put strikes on my dad’s account for me commenting on these videos to say that they are spam and scamming people.”

TikTok users who have gotten wise to the scheme have commented on videos calling out the scam. Like Yelichek, some comments say that TikTok responds to their reports of a video by saying it does not violate its community guidelines.

TikTok said in a statement that its community guidelines do, in fact, prohibit impersonation accounts and content that violates others’ intellectual property rights. Reports of copyright infringement concerns may require proof of ownership, including links to the original content and links to infringing content. Meta similarly stated its Instagram terms of use do not allow posting content that violates someone else’s intellectual property rights, including copyright and trademarks, and that violations can be reported on Instagram’s help page.

Ray, the 85-year-old jokester, said he tried reporting the videos to TikTok but got responses that the content he had flagged did not violate TikTok’s community standards. Like Yelichek, he said his comments on the videos alerting followers to the scam were frequently removed, and that he gets “no help from TikTok” and does not know how to further communicate with the company.

TikTok said in a statement that it continuously takes action against such copyright violations. It added that 94% of all content removed for violating community guidelines on fake engagement in the fourth quarter of last year was removed proactively rather than in response to reports.

Meanwhile, some users are catching on to this particular kind of sadness bait – with recent videos going viral alerting people to the fact that the posts begging for help with failing cat shelters, cow farms and other heart-wrenching fictions are not real.

Though awareness may spread, the impersonating videos remain available. In lieu of takedowns, Ray said he has decided to continue to make his videos because many of his 13,000 followers have told him they look forward to his posts each day.

“With everything that’s going on nowadays, people need to smile,” he said. “If my followers smile, even for a minute, they’ve forgotten their troubles for a minute. So that’s all I try to do – to make people smile. This is not going to stop me.”

Fraudsters are manipulating content of older adults as ‘sadness bait’ to push sales. The videos are going viral

‘It’s thieving’: impersonators steal elderly people’s TikToks to hawk mass-produced goods

Internet users have been told to change their passwords and upgrade their digital security after researchers claimed to have revealed the scale of sensitive information – 16bn login records – potentially available to cybercriminals.

Researchers at Cybernews, an online tech publication, said they had found 30 datasets stuffed with credentials harvested from malicious software known as “infostealers” and leaks.

The researchers said the datasets were exposed “only briefly” but amounted to 16bn login records, with an unspecified number of overlapping records – meaning it is difficult to say definitively how many accounts or people have been exposed.

Cybernews said the credentials could open access to services including Facebook, Apple and Google – although there had been no “centralised data breach” at those companies.

Bob Diachenko, the Ukrainian cybersecurity specialist behind the research, said the datasets had become temporarily available after being poorly stored on remote servers – before being removed again. Diachenko said he was able to download the files and would aim to contact individuals and companies that had been exposed.

“It will take some time of course because it is an enormous amount of data,” he said.

Diachenko said the information he had seen in infostealer logs included login URLs to Apple, Facebook and Google login pages. Apple and Facebook’s parent, Meta, have been contacted for comment.

A Google spokesperson said the data reported by Cybernews did not stem from a Google data breach – and recommended people use tools like Google’s password manager to protect their accounts.

Internet users are also able to check if their email has been compromised in a data breach by using the website haveibeenpwned.com. Cybernews said the information seen in the datasets followed a “clear structure: URL, followed by login details and a password”.

Diachenko said the data appeared to be “85% infostealers” and about 15% from historical data breaches such as a leak suffered by LinkedIn.

Experts said the research underlined the need to update passwords regularly and adopt tough security measures such as multifactor authentication – or combining a password with another form of verification such as a code texted from a phone. Other recommended measures include passkeys, a password-free method championed by Google and Facebook’s owner, Meta.

“While you’d be right to be startled at the huge volume of data exposed in this leak it’s important to note that there is no new threat here: this data will have already likely have been in circulation,” said Peter Mackenzie, the director of incident response and readiness at the cybersecurity firm Sophos.

Mackenzie said the research underlined the scale of data that can be accessed by online criminals.

“What we are understanding is the depth of information available to cybercriminals.”



He added: “It is an important reminder to everyone to take proactive steps to update passwords, use a password manager and employ multifactor authentication to avoid credential issues in the future.”

Toby Lewis, the global head of threat analysis at the cybersecurity firm Darktrace, said the data flagged in the research is hard to verify but infostealers – the malware reportedly behind the data theft – are “very much real and in use by bad actors”.

He said: “They don’t access a user’s account but instead scrape information from their browser cookies and metadata. If you’re following good practice of using password managers, turning on two-factor authentication and checking suspicious logins, this isn’t something you should be greatly worried about.”

Cybernews said none of the datasets have been reported previously barring one revealed in May with 184m records. It described the datasets as a “blueprint for mass exploitation” including “account takeover, identity theft, and highly targeted phishing”.

The researchers added: “The only silver lining here is that all of the datasets were exposed only briefly: long enough for researchers to uncover them, but not long enough to find who was controlling vast amounts of data.”

Alan Woodward, a professor of cybersecurity at Surrey University, said the news was a reminder to carry out “password spring cleaning”. He added: “The fact that everything seems to be breached eventually is why there is such a big push for zero trust security measures.”

Hacked credentials could give cybercriminals access to Facebook, Meta and Google accounts among others

Internet users advised to change passwords after 16bn logins exposed

View image in fullscreen Illustration: The Guardian/Charles Desmarais

With electric cars skyrocketing in popularity around the world – in 2024, 22% of new car sales worldwide were electric vehicles, compared with 18% in 2023 – a growing body of studies and an increasing number of people have found that they feel more motion sick riding in EVs than in traditional petrol or diesel cars. Anecdotes of feeling sick in the passenger or back seat of electric cars litter social media, as do questions from wary prospective buyers.

There is a scientific explanation behind why a person might feel more sick in an EV, though, according to multiple academic studies.

“Greater sickness in EVs can be attributed to a lack of previous experience, as both a driver and as a passenger, where the brain lacks accuracy in estimating the motion forces because it relies on previous experience in other types of cars,” said William Emond, a PhD student researching car sickness at the Université de Technologie de Belfort-Montbéliard in France.

Though EVs are becoming more popular, combustion cars still dominate. Riders have a longer history with gas cars and so have had more time to adapt to their specific cues. If a person has spent most of their life driving a combustion engine car, their brain anticipates acceleration after the rev of the engine, a warning that they are about to experience a change in speed. In a battery-powered car, the electric motor makes no such noises.

In addition to general unfamiliarity, research has found links between specific features common to electric vehicles and motion sickness. One 2024 study concluded that there were strong correlations between motion sickness severity and the seat vibrations of electric vehicles, while a 2020 study found that the lack of engine sound in an EV might be a major contributing factor to increased feelings of carsickness.

“If we are accustomed to traveling in non-EVs, we are used to understanding the car’s motion based on signals such as engine revs, engine vibrations, torque, etc. Yet, traveling in an EV for the first time is a new motion environment for the brain, which needs adaptation,” Emond explains.

Additionally, the regenerative braking technology used in EVs – where the motor converts the slowing car’s kinetic energy into electricity that then is stored in the battery – results in low-frequency deceleration, meaning that the vehicle slows down gradually and steadily, over a relatively longer period, rather than rapidly or in quick pulses. Such low-frequency deceleration tends to be associated with higher levels of motion sickness.

A 2024 study suggested the feature acted as one of the main triggers of motion sickness in electric vehicles. The study’s authors wrote: “Our results confirmed that higher levels of RB [regenerative braking] can induce MS [motion sickness].”

Motion sickness is thought to be caused by a mismatch between various sensory signals the brain simultaneously receives about the body’s movement. Specifically, it happens when the inner ear, which helps control balance, the eyes, and the body send conflicting information to the brain.

“Better knowledge on self-motion allows us to anticipate motion forces, which is crucial for motion sickness. Yet, when the motion forces as estimated or anticipated by the brain differ from what actually is experienced, then the brain interprets this ‘neural mismatch’ as a situation of conflict,” Emond said. “If this conflict persists over time, it may surpass a threshold for triggering autonomic reactions of the body such as symptoms apparent to ‘motion sickness’.”

Being able to anticipate the movement of a vehicle appears to be integral to the experience of motion sickness, which is why people who are driving a car don’t tend to experience symptoms. They know what is coming.

The interplay between anticipation of motion and actual motion experienced may be why EVs are associated with worse motion sickness, as these vehicles provide fewer clues regarding upcoming movements.

“When discovering a new motion environment, the brain needs to habituate because there is no knowledge of previous experience in such a context. This is, for example, why almost everyone becomes sick in zero-gravity environments,” Emond says.

As EV ownership continues to increase, some researchers are already looking into a solution for the cars’ specific type of nausea. Several research papers have suggested that motion sickness in autonomous EVs could be treated by using visual signals, such as interactive screens and ambient lighting, or vibrational cues to allow a passenger’s brain to anticipate motion changes, alleviating the lurching feeling in the back of an electric taxi.

An increasing number of people are experiencing motion sickness in EV and there is a scientific explanation as to why

Do electric vehicles make people more carsick?

A rash of recent assassinations have brought on congressional scrutiny and concern among law enforcement agencies who are wary of an age of political polarization turning deadly.

But experts say the violence is as much a byproduct of the times as it is the easy accessibility to DIY murder tradecraft, evident in some high-profile slayings of late.

So while the willingness to commit these acts has certainly increased, the tradecraft to pull them off has never been more obtainable.

“Political polarization, combined with the idea that one’s opponents are irredeemably evil and that there are no other legal avenues to create change, can lead to violence,” said Joshua Fisher-Birch, a terrorism analyst who closely tracks extremists across the political spectrum.

“There are several guides online for assassination, guerrilla warfare or similar violent acts, as well as counter-surveillance manuals shared by individuals in communication apps such as Telegram and online libraries.”

The proliferation of those kinds of resources have spilled into the mainstream and have given the average person the knowhow to access the types of instructions on popular apps that were once only available on dark web archives. Whether it’s downloading blueprints for and creating a 3D-printed gun, professionally tracking down targets, cooking up a bomb recipe or looking up ways to evade law enforcement once an act has been committed, a number of public attacks show so-called “lone wolves” using internet resources to plot their crimes.

For example, police say Vance Boelter, 57, charged with killing Democratic state representative Melissa Hortman and her husband at their home in Minnesota earlier this month, used a number of data brokering websites to amass not only a hitlist of “mostly or all Democrats” but their home addresses, too.

Court documents say Boelter was caught with “lists of internet-based people search engines including Truepeoplesearch, Spokeo, Pipl, Peoplefinders, Beenverified, Whitepages, Truthfinder, Intelius, Ownerly, USsearch and Peoplelooker” which can “aggregate data from various online and offline sources to provide the querying user details about the searched-for person, such as home addresses, phone numbers and information about family members”.

Extremists on the far right have also used similar resources to create kill lists of their enemies, and Fisher-Birch said some have become more “adept at doxing in recent years”. He noted that terrorist groups have gone about producing easily digestible manuals in pdf form then spreading them on Telegram, while other instructional materials are “commercially available books written by former members of militaries or intelligence services” they suggest reading.

Available online materials can also extend to sourcing weaponry. Fisher-Birch continued: “According to police, Luigi Mangione used a partially 3D-printed pistol to kill UnitedHealthcare CEO Brian Thompson.”

Mangione, who has amassed a cult-like following among anti-corporatists after the Manhattan assassination of Thompson in December, also allegedly managed to evade a manhunt for days with a stunning amount of counter-surveillance tradecraft – something other known attackers have achieved.

Still, to this day, the masked and hooded suspect seen planting pipe bombs at the RNC and DNC offices in Washington DC the night before the 6 January 2021 insurrection on Capitol Hill has yet to be caught. In the past, multiple law enforcement sources remarked how the surveillance footage shows the would-be bomber wearing the same prescribed disguise and gloves featured in terrorism manuals circulated on the far right.

Incidents like that suggest the ecosystem of DIY manualling has partly been seeded by extremist organizations on social media, which have then spread elsewhere. As early as 2018, the internationally designated neo-Nazi terrorist group the Base posted military tradecraft gleaned from US Marine Corps manuals and other sources instructing how to make a covert exit if you’re pursued by authorities.

Other jihadists groups, such as the Islamic State, have uploaded similar literature on encrypted forums they use to communicate with recruits.

skip past newsletter promotion Sign up to This Week in Trumpland Free newsletter A deep dive into the policies, controversies and oddities surrounding the Trump administration Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

“More generally, violent extremists groups put considerable time and effort into inciting followers to plot attacks and building robust online communities for information sharing and advice,” said Lucas Webber, a senior threat intelligence analyst at Tech Against Terrorism and a research fellow at the Soufan Center.

“[IS] and al-Qaida produce and disseminate manuals and videos teaching followers to construct explosive devices, detonators, suppressors and more.”

Coinciding with this surge in political violence is the growing willingness of nation-states to increase covert assassination operations on foreign soil, taking advantage of the moment. Iranian agents sponsored the attempted assassination of a dissident in New York by tapping criminal networks to carry it out, while Indian diplomats tried to disguise the murder of a Sikh activist in Canada as a gangland killing.

Of course, political assassinations and their attempts, without the clear aid of online tools, have also come to pass: a man stalked an event at a Jewish museum in DC to gun down Israeli diplomats in May, while then presidential candidate Donald Trump faced an attempted assassination last summer in the lead-up to the November vote – both attackers using over-the-counter firearms and limited tradecraft.

One thing does seem clear: authorities appear unable or unsure on how to counter the rising threat of lethal political violence.

The FBI declined to comment on the broad string of recent political assassinations and how the agency plans to handle them in the future.

“We have nothing to add to previous statements about the events you mention,” said a spokesperson in an email.

After Boelter was caught, a federal agent in Minnesota called his alleged murders “an appalling act of political violence that has no place in our country”.

People locators, 3D weapon blueprints, tactical planning – all accessible on the web for potential attackers or terrorists

Recent US political violence aided by DIY murder tradecraft available on internet

The UK’s grocery industry watchdog has launched an investigation into Amazon over allegations that the retail and technology company is failing to pay its suppliers on time.

The Groceries Code Adjudicator (GCA) said it had “reasonable grounds” to suspect that Amazon had breached a part of the groceries supply code of practice, which mandates that there should not be delays in payments made to suppliers.

The investigation into Amazon comes almost a year after the GCA told the online retailer it needed to take “swift and comprehensive action” to improve its compliance with industry rules designed to protect suppliers.

The GCA is responsible for regulating the relationships between the UK’s 14 largest grocery retailers – including the biggest supermarkets Tesco and Sainsbury’s – and their direct suppliers.

The investigation into Amazon’s grocery retail arm is its third since the GCA was established in 2013 after inquiries into Tesco and the Co-op. The watchdog has the power to fine a company up to 1% of turnover if it is found to have broken the groceries code.

The adjudicator, Mark White, said: “Delays in payment can significantly harm suppliers. The alleged delays could expose Amazon suppliers to excessive risk and unexpected costs, potentially affecting their ability to invest and innovate.”

The company sells groceries in the UK through its Amazon Fresh branded stores and online retail operation, as well the Whole Foods chain, which it bought in 2017 for $13.7bn (£10.2bn).

Suppliers raised more issues with Amazon than with other retailers, according to a survey for the GCA carried out in 2024.

After the warning to Amazon last July, the GCA said it had monitored the retailer’s actions and heard detailed evidence about suppliers’ experiences.

The GCA said on Friday it had grounds to suspect Amazon had breached paragraph 5 of the groceries code between 1 March 2022 and 20 June 2025, but would focus its investigations on the period from the start of 2024 to gain the best understanding of Amazon’s current practices.

The watchdog is calling on suppliers to send in evidence by 8 August. It added that all responses would remain confidential.

skip past newsletter promotion Sign up to Headlines UK Free newsletter Get the day’s headlines and highlights emailed direct to you every morning Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The GCA said it had also received information about other issues at Amazon, adding that it would “not hesitate” to launch further investigations if required.

It has previously reprimanded Tesco over its treatment of suppliers after a year-long investigation but said it was not able to fine the supermarket chain because the alleged offences were committed before the government handed the GCA extra powers in 2015, allowing it to fine companies in 2015. It also investigated Co-op in 2018.

Amazon said it took “the groceries supply code of practice incredibly seriously” and added that it would “cooperate fully with the adjudicator”.

The spokesperson said: “While we are disappointed with this decision, we welcome the opportunity to further demonstrate our ongoing compliance with this particular section of the code.



“We have already made significant improvements to our grocery supplier experience, including to payment practices, with supplier contacts on this reducing year-on-year.”



“We will continue to listen and work with our grocery suppliers as we roll out further changes.”

Regulator says it has ‘reasonable grounds’ to suspect company breached groceries supply code of practice

Amazon under UK investigation over alleged failure to pay suppliers on time

Garmin’s new top running watch, the Forerunner 970, has very big shoes to fill as it attempts to replace one of the best training and race companions available. Can a built-in torch, a software revamp and voice control really make a difference?

The Guardian’s journalism is independent. We will earn a commission if you buy something through an affiliate link. Learn more.

The new top-of-the-line Forerunner takes the body of the outgoing Forerunner 965 and squeezes in a much brighter display, useful new running analytics and more of the advanced tech from Garmin’s flagship adventure watch the Fenix 8.

These upgrades come at a steep cost of £630 (€750/$750/A$1,399) – £30 more than its predecessor – placing it right at the top of the running and triathlon watch pile, although less than the £780 Fenix 8.

View image in fullscreen The Forerunner 970 is a large but not enormous watch with a relatively low profile that slips under shirt cuffs and stays put during exercise. Photograph: Samuel Gibbs/The Guardian

The 970 is about the same size as the outgoing 965 with a 47mm case and a beautiful, crisp and very bright 1.4in OLED screen. The touchscreen is covered in super-hard sapphire glass similar to luxury watches, while the titanium bezel finishes off the polycarbonate body in a choice of three colours.

Quite a lot of the upgrades are trickle-downs from the Fenix 8 and make the 970 a better everyday smartwatch. It has Garmin’s new offline voice control system, which allows you to quickly set timers and alarms, access settings or start activities. The watch also connects to your phone’s voice assistant and takes calls on your wrist via Bluetooth.

A revamped interface speeds up access to notifications from your smartphone by swiping down from the top of the screen. With an iPhone you can view and dismiss text-only notifications but connected to an Android phone you can also see images in notifications and directly reply to them from the watch. The 970 has Garmin Pay for contactless payments, although bank support is limited, and can control music on your phone or download playlists from Spotify, Amazon Music, YouTube Music and others for phone-free music on runs.

View image in fullscreen Once you have a torch built into your watch you’ll never want to be without it. I use it at least five times a day. Photograph: Samuel Gibbs/The Guardian

The best new feature is the LED torch built into the 970’s top edge. It was invaluable on the Fenix 8 and is my favourite new addition to the Forerunner. It is bright enough to light your way on the street at night or find things buried in dark cupboards but can be turned a dim red to avoid waking everyone at home. It can also be used as a strobe light for running to help keep you visible at night.

The battery lasts about six days with general smartwatch usage, including having the screen on all the time, all-day and night monitoring of health, plenty of notifications and copious use of the torch. The screen has automatic brightness but turning it down one notch in settings, which was still plenty bright enough to see outdoors, added a couple of days to the battery life. Turning the always-on display setting off extended it further to about 12 to 15 days.

Specifications

Screen: 1.4in AMOLED (454x454)

Case size: 47mm

Case thickness: 13.2mm

Band size: standard 22mm

Weight: 56g

Storage: 32GB

Water resistance: 50 metres (5ATM)

Sensors: GNSS (Multiband GPS, Glonass, Galileo), compass, thermometer, heart rate, pulse Ox

Connectivity: Bluetooth, ANT+, wifi

Running and activity tracking

View image in fullscreen The watch charges via a USB-C cable that slots in next to the sensor cluster on the back, reaching 70% in about 40 minutes and full in 83 minutes. Photograph: Samuel Gibbs/The Guardian

Its predecessor was a fantastic running watch filled to the brim with metrics, helpful analysis and buckets of customisation options, on which the 970 only builds. The screen is large enough to be able to clearly see up to eight data fields on screen at once. Maps look particularly good and are easy to use with touch.

It has the latest dual-band GPS, while Garmin’s algorithms consistently have higher tracking accuracy than its rivals, even with similar systems. The new Gen 5 Elevate heart rate sensor on the back improves pulse monitoring in tricky conditions, and provides ECG (arrhythmia) readings.

The 970 has Garmin’s suite of industry-leading fitness, recovery and training metrics, which are joined by a few new and interesting statistics, including two that attempt to help you prevent injury.

Impact load quantifies how hard a run is on your body based on its intensity and difficulty compared with an easy, flat run at slower speeds. One fast, hard 7km run was rated as equivalent to a gentler 12km run, which felt about right in my feet and legs and made me consider taking a longer recovery time before the next workout.

View image in fullscreen Impact load helps you avoid overdoing your training by estimating the impact of your runs on your body. Photograph: Samuel Gibbs/The Guardian

In addition, the new running tolerance feature tracks your mileage over a seven-day period and advises how much more you can run without increasing your chance of injury. Many runners, including myself, have injured themselves when ramping up their weekly distance too fast when training for a race, which this new stat is an attempt to avoid by giving you suggested guard rails.

The 970 also has a new running economy feature that tracks efficiency of your form, including how much speed you lose as your foot hits the ground, but it relies on Garmin’s latest heart rate monitor strap, the HRM 600 – a £150 separate purchase.

Running battery life is a solid 11-plus hours with its highest accuracy settings and listening to offline music via Bluetooth headphones, or about 16 hours without music. Turning down the screen brightness a bit added several hours to the running battery life, while reducing the GPS accuracy mode can last up to 26 hours.

Solid general health monitoring

View image in fullscreen The Forerunner looks and feels premium with flashes of colour on the body and strap but is obviously a sports watch, so may not be suitable for all occasions. Photograph: Samuel Gibbs/The Guardian

The Garmin isn’t entirely about running, triathlon and its 30-plus sport tracking features. It also has a comprehensive suite of general health monitoring tools, including good sleep, activity, stress, women’s health and heart health tracking rivalling an Apple Watch or similar.

Most of Garmin’s most advanced training tools also monitor your recovery from exercise during the rest of the day and night, advising you in the morning and during the day how your body is doing. It has a built-in sleep coach, a running or triathlon coach and various advisers for activity, suggesting when to do a hard workout and when to take it easy. The daily suggested workouts are dynamic and based on your sleep and recovery, so it will never prompt you to do a hardcore workout when you’ve had a terrible night. These automatic workouts can be replaced by a coaching plan, either using Garmin’s solid tools or third-party ones placed on a calendar before a race.

Sustainability

The watch is generally repairable with options available via support. The battery is rated to maintain at least 90% of its original capacity after two years of weekly charging. The watch does not contain any recycled materials. Garmin guarantees security updates until at least 21 May 2027 but typically supports its devices far longer. It offers recycling schemes on new purchases.

Price

The Garmin Forerunner 970 costs £629.99 (€749.99/$749.99/A$1,399).

For comparison, the Garmin Fenix 8 costs from £780, the Forerunner 570 costs £460, the Garmin Forerunner 965 costs £499.99, the Apple Watch Ultra 2 costs £799, the Coros Pace Pro costs £349, the Suunto Race S costs £299 and the Polar Vantage V3 costs £519.

Verdict

Garmin continues to set the bar for running watches with the Forerunner 970. It isn’t a dramatic leap over the outgoing Forerunner 965, instead adding a few bits to the already excellent formula.

The screen is brighter, covered in scratch-resistant sapphire and ringed by a titanium bezel, which gives it a premium look and feel alongside a more modern and responsive interface. The added bells and whistles of voice control and faster access to notifications make using it as a smartwatch alternative much easier. Though wearing it is still a statement about your sporty priorities compared with an Apple or Pixel Watch.

The upgraded heart rate sensor helps keep things locked during more difficult exercises and adds ECG readings for more comprehensive heart health tracking. But it is the built-in torch that is the best addition for daily life. Every watch should have one.

Meanwhile, the new impact load and running tolerance features could be very useful for avoiding strain and injury, adding to the already excellent training and recovery tracking. Plus it has market-leading running accuracy and detailed onboard maps for routes or if you get lost.

If you want a premium running and triathlon watch with all the bells and whistles, the Forerunner 970 is the best you can get. It just comes at a very high cost.

Pros: super bright OLED screen, built-in torch, phone and offline voice control, Garmin Pay, extensive tracking and recovery analysis for running and many other sports, full offline mapping, offline Spotify, buttons and touch, most accurate GPS, ECG. Cons: expensive, limited Garmin Pay bank support, still limited smartwatch features compared with Apple/Google/Samsung watches, battery life shorter than LCD rivals.

Premium sports tracker adds built-in torch, smartwatch and accuracy upgrades, plus useful new training tools, but costs far more than rivals

Garmin Forerunner 970 review: the new benchmark for running watches

Many Uber drivers are earning “substantially less” an hour since the ride hailing app introduced a “dynamic pricing” algorithm in 2023 that coincided with the company taking a significantly higher share of fares, research has revealed.

The findings are in a study released on Thursday by academics at the University of Oxford. They analysed data provided by 258 UK Uber drivers responsible for 1.5m trips.

Having initially taken a fixed 20% cut of the UK fares charged, which subsequently rose to 25%, Uber introduced dynamic pricing in 2023, an algorithm that variably sets pay for drivers and fares for passengers. It is a later iteration of Uber’s “surge pricing” that increased fares during periods of peak demand.

Uber is now claiming a cut, or “take rate”, of 29% of a fare, rising to more than 50% in some cases, the researchers found.

Unions criticised the move when it was made in 2023, claiming there was no transparency and that the technology “could push down working conditions by targeting drivers based on their willingness and ability to accept lower fares”.

The Oxford research said: “Post-dynamic pricing, Uber’s passengers now pay higher prices, but the drivers are not better off.”

The paper, which was published in partnership with the non-profit gig worker organisation Worker Info Exchange (WIE), concluded: “Our findings suggest that post-dynamic pricing, many aspects of Uber drivers’ jobs have gotten worse. Average pay per hour on the app is stagnant, and is lower in real terms in the year following the introduction of dynamic pricing.

“Uber’s median take rate per driver has increased from 25% to 29%, and on some trips the take rate is over 50%. Furthermore, the higher take rates are concentrated among higher-fare trips, which explains how Uber can extract an additional 38% [income] from its driver’s labour on average … Many drivers are earning substantially less per hour.”

The findings follow a series of controversies to have engulfed the technology firm, including a 2021 UK supreme court ruling that Uber drivers are entitled to the minimum wage and paid holidays, as well as the 2022 release of the Uber files, a global investigation that revealed how the company duped police and regulators, and secretly lobbied governments across the world.

After the release of the Uber files, Jill Hazelbaker, Uber’s senior vice-president of public affairs, said: “We have not and will not make excuses for past behaviour that is clearly not in line with our present values. Instead, we ask the public to judge us by what we’ve done over the last five years and what we will do in the years to come.”

The Oxford research added that drivers’ average hourly pay was £29.46, using an Uber definition, or £15.98 if counting waiting time when they made themselves available to pick up passengers. Neither average takes into account costs including vehicle maintenance, insurance or fuel.

skip past newsletter promotion Sign up to Business Today Free daily newsletter Get set for the working day – we'll point you to all the business news and analysis you need every morning Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Uber said it did “not recognise the figures in this report”, adding: “Every driver is guaranteed to earn at least the national living wage.”

One interviewee in the study said it was only when passengers volunteered the fares they paid in conversations with drivers that “you discover they [Uber] are robbing us and the customer”.

An Uber spokesperson said: “Uber drivers in the UK took home over £1bn in earnings between January and March of this year, which is up on the year before. Drivers choose to drive with Uber because we offer total flexibility on when they work and provide full transparency over the trips they accept.

“All drivers receive a weekly summary of their earnings, which includes a clear breakdown of what Uber and the driver received from trips. We are proud that thousands of drivers continue to make the positive choice to work on Uber as passenger demand and trips continue to grow.”

Introduction of ‘dynamic pricing’ also coincided with company raising trip prices, researchers finds

UK Uber drivers’ earnings cut after secretive algorithm changes

More than a decade after being one of 19 Uber drivers who took the company to court in 2015, Abdurzak Hadi continues to drive for – and fight with – the ride hailing app.

The group won their claim confirming their entitlement to the legal minimum wage – but the Silicon Valley company’s insistence that its drivers were self-employed contractors meant the case went all the way to the supreme court. In 2021, Hadi and friends won there too.

If that sounds as if the British legal system left the former Somalian refugee in the driving seat, he argues that life for Uber workers is now as precarious as ever.

On Thursday, academics at the University of Oxford – in conjunction with the non-profit gig worker organisation Worker Info Exchange (WIE) – launched a report analysing a mass of data relating to 1.5m trips provided by 258 UK Uber drivers, who had used privacy legislation to extract their personal data from the ride-hailing app.

The study gave a rare opportunity to study the workings of Uber’s technology and produced some eye-catching findings.

It found that many Uber drivers have earned “substantially less” an hour since the app introduced a “dynamic pricing” algorithm in 2023, which is said to adjust trip prices in real time based on a number of factors.

These include time, distance, the number of available drivers in the area, the passenger demand, traffic and the weather. The report found that these earnings drops coincided with the company taking a significantly higher share of fares.

The study unearthed further data suggesting that drivers such as Hadi are experiencing less and less control over their working lives. They described their days as being controlled by increasingly sophisticated pieces of computer code, which left them unclear how much Uber would take in fees on discrete jobs.

“[The old system] was clear, transparent,” Hadi told the Guardian and ITV News. “You can calculate, you can see. Say, for example here it says about eight miles, so I know eight miles plus how long it took me, plus the starting fare, minus Uber’s fee, which is 20%. Even when they increased it to 25%, I would exactly know how much. Exactly.”

The new system has resulted in Uber taking a variable cut, or “take rate”, of 29% of a fare on average, rising to more than 50% in some cases, the University of Oxford researchers found. The study also found that Uber’s take rate increased on higher value rides – something the company has denied.

The 29% figure appears to chime with disclosures within Uber’s latest quarterly results figures, which show that the company made $1.2bn (£890m) of income from its operations in the first three months of this year.

Meanwhile, the WIE estimates that UK Uber drivers lost out on $1.6bn in pay, as a result of Uber increasing its share of the fare, during the 12 months to March 2025.

Uber said the UK take rate and lost earnings figures are inaccurate, and that its take rate had remained “steady” at 25%.

The company added: “The Uber app reviews real-time information to provide the best price to appeal to the drivers in the area, helping to minimise waiting times for customers and maximise earnings. Drivers are shown their earnings for the trip before they decide whether to accept.”

But as Uber grows more confident in its calculations, those transporting passengers say they are becoming less so.

A driver’s livelihood depends on their ability to guess what kinds of trips they will get at particular times and places, and how much those trips will pay.

However, the University of Oxford study stated that “drivers frequently complained about the unpredictability of pay post-dynamic pricing”.

It continued: “Any tacit knowledge drivers have built up over years about how much pay a given trip is likely to yield may no longer help them … the predictability of pay drastically changed after dynamic pricing was introduced.”

A company spokesperson said: “Uber drivers in the UK took home over £1bn in earnings between January and March of this year, which is up on the year before.

“Drivers choose to drive with Uber because we offer total flexibility on when they work and provide full transparency over the trips they accept.

“All drivers receive a weekly summary of their earnings, which includes a clear breakdown of what Uber and the driver received from trips.

“We are proud that thousands of drivers continue to make the positive choice to work on Uber as passenger demand and trips continue to grow.”

Ride-hailing app’s ‘take rate’ – or cut – sometimes reaches more than 50% since introduction of dynamic pricing

Rough ride: how Uber quietly took more of your fare with its algorithm change

Donald Trump has signed an executive order to delay the ban or sale of TikTok for the third time. The order gives the Chinese-owned social media company another 90 days to find a buyer or be banned in the US.

“I’ve just signed the Executive Order extending the Deadline for the TikTok closing for 90 days (September 17, 2025),” the president said in a Truth Social post.

Trump’s first executive order giving TikTok a reprieve came on his first day in office – just three days after the supreme court ruled to uphold the ban. Trump issued the second executive order in April. The deadline for the sale or ban was then set for 19 June. Now, TikTok has until September.

Mark Warner, a Democratic senator and the vice-chair of the Senate intelligence committee, accused Trump of sidestepping the law with an executive order.

“Once again, the Trump administration is flouting the law and ignoring its own national security findings about the risks posed by a PRC-controlled TikTok,” Warner said in a statement.

TikTok is a tremendously popular social media app with 170 million users in the US. ByteDance, TikTok’s parent company, said in April that it had been in discussion with the US government regarding a solution for the app. It added that any agreement “will be subject to approval under Chinese law”.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Before Trump issued his second executive order on TikTok, a possible deal was reportedly in place. But, according to Reuters, China put that deal on hold after Trump announced sweeping tariffs against the country.

At the time, Trump indicated he was using tariffs for leverage in the deal. “We have a situation with TikTok where China will probably say we’ll approve a deal, but will you do something on the tariffs. The tariffs give us great power to negotiate,” Trump said.

The idea of banning TikTok originated with Trump in 2020, who said the Chinese-owned app posed a danger to national security. It quickly became a bipartisan issue and Congress overwhelmingly voted to ban the app last year, which faced a legal challenge but was ultimately affirmed by the supreme court. The original deadline for the TikTok ban was on 19 January.

Trump switched his stance on TikTok after joining the app while campaigning for president last year, amassing nearly 15 million followers and hosting the TikTok CEO, Shou Zi Chew, at Trump’s Mar-a-Lago estate in Florida. Chew also attended Trump’s inauguration.

Several US investors have expressed interest in buying the social media app, which is probably worth tens of billions of dollars. Among those who have reportedly thrown in bids are a consortium led by the software giant Oracle, asset manager Blackstone, Amazon, Walmart, billionaire Frank McCourt, a crypto foundation and the founder of the adult website OnlyFans.

TikTok, ByteDance and the White House did not return requests for comment.

Order gives the Chinese-owned social media app, which has 170 million users in the US, until September to find a buyer

Trump signs order granting TikTok a third reprieve from US ban

The Meta chief executive, Mark Zuckerberg, called it “the most intelligent AI assistant that you can freely use”. But Barry Smethurst, 41, a record shop worker trying to travel by rail from Saddleworth to Manchester Piccadilly, did not agree.

Waiting on the platform for a morning train that was nowhere to be seen, he asked Meta’s WhatsApp AI assistant for a contact number for TransPennine Express. The chatbot confidently sent him a mobile phone number for customer services, but it turned out to be the private number of a completely unconnected WhatsApp user 170 miles away in Oxfordshire.

It was the beginning of a bizarre exchange of the kind more and more people are having with AI systems, in which chatbots try to negotiate their way out of trouble, deflect attention from their mistakes and contradict themselves, all in an attempt to continue to appear useful.

Smethurst’s joust with Meta’s AI began when he told it the TransPennine helpline it had dug up looked like a private number. When he asked what this number was, the AI quickly said it shouldn’t have shared it and tried to change the subject, saying: “Let’s focus on finding the right info for your TransPennine Express query!”

But Smethurst wanted to know why it had shared the private number, which it turned out belonged to James Gray, 44, a property industry executive from Oxfordshire who is also a WhatsApp user and whose number appears on his company website.

The AI explained vaguely it was generated “based on patterns” and promised to “strive to do better in the future” by admitting when it didn’t know an answer. But it then falsely described the number as “fictional” and not “associated with anyone”. When Smethurst challenged that, it admitted: “You’re right,” and said it may have been “mistakenly pulled from a database”.

Asked which database, it muddied the waters further by contradicting itself again saying: “I didn’t pull the number from a database. I generated a string of digits that fit the format of a UK mobile number but it wasn’t based on any real data on contacts.”

With the conversation going round in circles, Smethurst said he did not believe the chatbot and told it: “Just giving a random number to someone is an insane thing for an AI to do.”

“It’s terrifying,” Smethurst said, after he raised a complaint with Meta. “If they made up the number, that’s more acceptable, but the overreach of taking an incorrect number from some database it has access to is particularly worrying.”

Gray said he had thankfully not received calls from people trying to reach TransPennine Express, but said: “If it’s generating my number could it generate my bank details?”

Asked about Zuckerberg’s claim that the AI was “the most intelligent”, Gray said: “That has definitely been thrown into doubt in this instance.”

Developers working with OpenAI chatbot technology recently shared examples of “systemic deception behaviour masked as helpfulness” and a tendency to “say whatever it needs to to appear competent” as a result of chatbots being programmed to reduce “user friction”.

In March, a Norwegian man filed a complaint after he asked OpenAI’s ChatGPT for information about himself and was confidently told that he was in jail for murdering two of his children, which was false.

And earlier this month a writer who asked ChatGPT to help her pitch her work to a literary agent revealed how after lengthy flattering remarks about her “stunning” and “intellectually agile” work, the chatbot was caught out lying that it had read the writing samples she uploaded when it hadn’t fully and had made up quotes from her work. It even admitted it was “not just a technical issue – it’s a serious ethical failure”.

Referring to Smethurst’s case, Mike Stanhope, the managing director of the law firm Carruthers and Jackson, said: “This is a fascinating example of AI gone wrong. If the engineers at Meta are designing ‘white lie’ tendencies into their AI, the public need to be informed, even if the intention of the feature is to minimise harm. If this behaviour is novel, uncommon, or not explicitly designed, this raises even more questions around what safeguards are in place and just how predictable we can force an AI’s behaviour to be.”

Meta said that its AI may return inaccurate outputs, and that it was working to make its models better.

“Meta AI is trained on a combination of licensed and publicly available datasets, not on the phone numbers people use to register for WhatsApp or their private conversations,” a spokesperson said. “A quick online search shows the phone number mistakenly provided by Meta AI is both publicly available and shares the same first five digits as the TransPennine Express customer service number.”

A spokesperson for OpenAI said: “Addressing hallucinations across all our models is an ongoing area of research. In addition to informing users that ChatGPT can make mistakes, we’re continuously working to improve the accuracy and reliability of our models through a variety of methods.”

Chatbot tries to change subject after serving up unrelated user’s mobile to man asking for rail firm helpline

‘It’s terrifying’: WhatsApp AI helper mistakenly shares user’s number

The boss of OpenAI has claimed that Mark Zuckerberg’s Meta has tried to poach his top artificial intelligence experts with “crazy” signing bonuses of $100m (£74m), as the scramble for talent in the booming sector intensifies.

Sam Altman spoke about the offers in a podcast on Tuesday. They have not been confirmed by Meta. OpenAI, the company that developed ChatGPT, said it had nothing to add beyond its chief executive’s comments.

“They started making these giant offers to a lot of people on our team – $100m signing bonuses, more than that comp [compensation] per year,” Altman told the Uncapped podcast, which is presented by his brother, Jack. “It is crazy. I’m really happy that, at least so far, none of our best people have decided to take them up on that.”

He said: “I think the strategy of a tonne of upfront, guaranteed comp, and that being the reason you tell someone to join … the degree to which they’re focusing on that, and not the work and not the mission – I don’t think that’s going to set up a great culture.”

Meta last week launched a $15bn drive towards computerised “super-intelligence” – a type of AI that can perform better than humans at all tasks. The company bought a large stake in the $29bn startup Scale AI, set up by the programmer Alexandr Wang, 28, who joined Meta as part of the deal.

Last week, a Silicon Valley venture capitalist, Deedy Das, tweeted: “The AI talent wars are absolutely ridiculous”. Das, a principal at Menlo Ventures, said Meta had been losing AI candidates to rivals despite offering $2m-a-year salaries.

Another report last month found that Anthropic, an AI company backed by Amazon and Google and set up by engineers who left Altman’s company was “siphoning top talent from two of its biggest rivals: OpenAI and DeepMind”.

The scramble to recruit the best developers comes amid rapid advances in AI technology and a race to achieve human-level AI capacity – known as artificial general intelligence. The spending on hardware is greater still, with recent estimates from the Carlyle Group, reported by Bloomberg, that $1.8tn could be spent on computing power by 2030. That is more than the annual gross domestic product of Australia.

Some tech firms are buying whole companies to lock in top talent, as seen in part with Meta’s Scale AI deal and Google spending $2.7bn last year on Character.AI, which was founded by the leading AI researcher Noam Shazeer. He co-wrote the 2017 research paper Attention is all you Need, which is considered a seminal contribution to the current wave of large language model AI systems.

While Meta was founded as a social media company and OpenAI as non-profit – becoming a for-profit business last year – the two are now rivals. Altman told his brother’s podcast that he did not feel Meta would succeed in it’s AI push, adding: “I don’t think they’re a company that’s great at innovation.”

He said he had once heard Zuckerberg say that it had seemed rational for Google to try to develop a social media function in the early days of Facebook, but “it was clear to people at Facebook that that was not going to work”.

skip past newsletter promotion Sign up to Business Today Free daily newsletter Get set for the working day – we'll point you to all the business news and analysis you need every morning Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

“I feel a little bit similar here,” Altman added.

Despite the huge investments in the sector, Altman suggested the result could be “we build legitimate super intelligence, and it doesn’t make the world much better [and] doesn’t change things as much as it sounds like it should”.

“The fact that you can have this thing do this amazing stuff for you, and you kind of live your life the same way you did two years ago,” he said.

“The thing that I think will be the most impactful in that five to 10-year timeframe is AI will actually discover new science. This is a crazy claim to make, but I think it is true, and if it is correct, then over time I think that will dwarf everything else [AI has achieved].”

Sam Altman describes offer from Mark Zuckerberg’s company as ‘crazy’, as scramble for talent intensifies

OpenAI boss accuses Meta of trying to poach staff with $100m sign-on bonuses

The boss of Amazon has told white collar staff at the e-commerce company their jobs could be taken by artificial intelligence in the next few years.

Andrew Jassy told employees that AI agents – tools that carry out tasks autonomously – and generative AI systems such as chatbots would require fewer employees in certain areas.

“As we roll out more generative AI and agents, it should change the way our work is done,” he said in a memo to staff. “We will need fewer people doing some of the jobs that are being done today, and more people doing other types of jobs.

“It’s hard to know exactly where this nets out over time, but in the next few years, we expect that this will reduce our total corporate workforce.”

Amazon employs 1.5 million people worldwide, with about 350,000 working in corporate jobs such as software engineering and marketing.

At the weekend the chief executive of the UK telecoms company BT said advances in AI could lead to deeper job cuts at the company, while Dario Amodei, the chief executive of the AI company Anthropic, said last month AI could wipe out half of all entry-level office jobs.

Jassy said in the near future there would be billions of AI agents working across companies and in people’s daily lives.

“There will be billions of these agents, across every company and in every imaginable field. There will also be agents that routinely do things for you outside of work, from shopping to travel to daily chores and tasks. Many of these agents have yet to be built, but make no mistake, they’re coming, and coming fast,” he said.

Jassy ended the memo by urging employees to be “curious about AI” and to “educate yourself” in the technology and take training courses.

“Those who embrace this change, become conversant in AI, help us build and improve our AI capabilities internally and deliver for customers, will be well positioned to have high impact and help us reinvent the company,” he said.

skip past newsletter promotion Sign up to Business Today Free daily newsletter Get set for the working day – we'll point you to all the business news and analysis you need every morning Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The Organisation for Economic Co-operation and Development – an influential international policy organisation – has estimated the technology could trigger job losses in skilled white-collar professions such as law, medicine and finance. The International Monetary Fund has calculated 60% of jobs in advanced economies such as the US and UK are exposed to AI and half of these jobs may be negatively affected.

However, the Tony Blair Institute, which has called for widespread adoption of AI in the public and private sectors, has said the technology could displace up to 3m private sector jobs in the UK but the net loss will be mitigated by the technology creating new roles.

Andrew Jassy tells white collar workers that such technology means fewer people will be needed for some jobs

Amazon boss tells staff AI means their jobs are at risk in coming years

An Israel-linked hacking group has claimed responsibility for a $90m (£67m) heist on an Iranian cryptocurrency exchange.

The group known as Gonjeshke Darande, Farsi for Predatory Sparrow, said on Wednesday it had hacked the Nobitex exchange, a day after claiming it had destroyed data at Iran’s state-owned Bank Sepah.

Elliptic, a consultancy specialising in crypto-related crime, said it had so far identified more than $90m in cryptocurrency sent from Nobitex crypto wallets to hacker addresses.

The hackers appear to have in effect “burned” those funds, rendering them inaccessible by storing them in “vanity addresses” for which they do not have the cryptographic keys, Elliptic said.

Tom Robinson, Elliptic’s co-founder, told the Guardian it would take current computer technology “billions of years” to create the cryptographic key pairs that match the vanity addresses.

The funds are being held in addresses containing some variation of the term “F*ckIRGCterrorists”. In a post on X, Predatory Sparrow said it had targeted Nobitex and would release its source code and “internal information”.

Predatory Sparrow is regularly described in Israeli media as being Israel-linked, although there has been no official confirmation of the hackers’ identity or their nationality.

“Although there is no confirmation yet that the funds were moved by Predatory Sparrow, the hack appears to be motivated by the recent escalation of tensions between Israel and Iran,” Elliptic said.

Rafe Pilling, the director of threat intelligence at the cybersecurity firm Sophos, said there was no firm evidence linking Predatory Sparrow to a particular state, but it had the characteristics of a government-backed group.

“It bears all the hallmarks of a false persona used by a government-sponsored threat group to conduct disruptive operations against targets linked to illicit Iranian revenue generation, logistical entities, transport infrastructure and other strategic sectors,” he said.

He added: “While we don’t expect to find strong technical links between Israel and Predatory Sparrow, the actions of the group align strongly with Israel’s regional priorities you’d be hard pushed to find another candidate country in the region with the capability to perform these attacks.”

Nobitex said on X it had experienced a “security incident” and was “actively working on implementing a secure and efficient recovery plan”.

Predatory Sparrow claimed in a post on X that it had “destroyed the data” of Bank Sepah and accused the bank of financing the Iranian military. Bank Sepah’s international branch in London has been approached for comment.

Meanwhile, companies tracking global internet activity have reported a near-total internet blackout in Iran, Cloudflare told the Guardian that traffic volumes were 98% below where they were at the same time last week.

However, hackers do not appear to have been the cause of the shutoff. An Iranian government spokesperson, Fatemeh Mohajerani, said this week that internet access had been slowed down to “maintain the network’s stability” and to ward of cyberattacks.

Hackers known as Predatory Sparrow claim responsibility for rendering Nobitex exchange funds inaccessible

Israel-linked group hacks Iranian cryptocurrency exchange in $90m heist

Up to seven out of 10 streams of artificial intelligence-generated music on the Deezer platform are fraudulent, according to the French streaming platform.

The company said AI-made music accounts for just 0.5% of streams on the music streaming platform but its analysis shows that fraudsters are behind up to 70% of those streams.

AI-generated music is a growing problem on streaming platforms. Fraudsters typically generate revenue on platforms such as Deezer by using bots to “listen” to AI-generated songs – and take the subsequent royalty payments, which become sizeable once spread across multiple tracks.

The tactic aims to evade detection measures triggered by vast listening numbers for a small amount of bogus tracks.

Thibault Roucou, the director of royalties and reporting at the Paris-based platform, said the manipulation of AI-generated music was an bid to “get some money from royalties”.

“As long as there is money [in fraudulent streaming] there will be efforts, unfortunately, to try to get a profit from it,” he said. “That’s why we’re investing in fighting it, because we know it’s not going away and we need to be one step ahead every time.”

Deezer deploys a tool it says can detect 100% AI-made content from the most prolific AI music models such as Suno and Udio.

Deezer said the AI-generated music streamed by fraudsters ranged from fake pop and rap music to artificial mood tracks. The platform blocks royalty payments for streams it has identified as fraudulent.

Deezer revealed in April that AI-generated music now represents 18% of all uploads to its platform, or 20,000 tracks per day. The company said it was removing all fully AI-generated content from its algorithmic recommendations. Deezer has more than 10 million subscribers worldwide, while market leader Spotify has 268 million.

Roucou said he did not know who was behind the fraudulent streams but the perpetrators appeared to be “organised”. The global streaming market was worth $20.4bn last year, according to trade body the IFPI, representing a sizeable target for fraudsters.

In its latest global music report, the IFPI said fraudulent streaming steals money that “should be going to legitimate artists”. It said generative AI had “significantly exacerbated” the problem.

Last year, a musician in the US, Michael Smith, was charged in connection with a scheme to create hundreds of thousands of AI-generated songs and stream them billions of times, obtaining $10m in royalty payments.

According to the French streaming platform’s analysis, fraudsters use bots to listen to AI-generated music and take the royalties

Up to 70% of streams of AI-generated music on Deezer are fraudulent, says report

When octopuses extend their eight arms into hidden nooks and crannies in search of a meal, they are not just feeling around in the dark for their food. They are tasting their prey, and with even more sensory sophistication than scientists had already imagined.

Researchers reported on Tuesday in the journal Cell that octopus arms are fine-tuned to “eavesdrop into the microbial world,” detecting microbiomes on the surfaces around them and deriving information from them, said Rebecka Sepela, a molecular biologist at Harvard and an author of the new study.

Where octopus eyes cannot see, their arms can go to identify prey and make sense of their surroundings. Scientists knew that those eight arms (not tentacles) sense whether their eggs are healthy or need to be pruned. And the hundreds of suckers on each arm have over 10,000 chemotactile sensory receptors each, working with 500 million neurons to pick up that information and relay it throughout the nervous system.

Yet, what exactly the octopus is tasting by probing and prodding — and how its arms can distinguish, say, a rock from an egg, a healthy egg in its clutch from a sick one or a crab that’s safe to eat from a rotting, toxic one — has long baffled scientists.

Scientists discovered that octopuses use their limbs to sample the microbiomes on the surfaces they touch.

Octopuses’ 8 Arms Snoop on Microbiomes

But Mr. Herbert holds far more in his head alone than the sum total of all that published research.

“I said, ‘That stupid little map right there, throw that in the garbage,’” Mr. Herbert recalled in a recording of a launch event for Dr. Cannon’s new book on Northern Dene star knowledge that was hosted by the Tanana Chiefs Conference, a consortium of Alaska Native communities.

The book, “In the Footsteps of the Traveller,” grew from that first meeting with Mr. Herbert and replaces earlier scholarly condescension with a clearer picture of a huge, ancient and intricate astronomical system shared by Elders across more than 750 miles of subarctic landscape. Alongside Mr. Herbert, some 65 Indigenous knowledge holders contributed to the book. More than a third have passed away since Dr. Cannon began the research.

One of the most central features of the regional astronomical system is a single figure who straddles the entire sky: In Gwich’in, he is Yahdii.

“We may have called him different names, but we all have the same story of the man in the sky,” Fred Sangris, of the Yellowknives Dene First Nation in Canada, wrote in the volume’s foreword. “As a chief, Elder, and contributor of this book, I give this work my sincere blessing and hope that many benefit from the wisdom in its pages.”

At the beginning, though, it was clear as he sat awkwardly at the kitchen table that Dr. Cannon, now an Indigenous studies scholar at the University of Alaska Fairbanks, had much to learn. Including how to learn. Dr. Cannon would find that such knowledge could not be acquired with traditional Western academic inquiry. Instead, he would need to enter a world with its own thought, methods and rules.

A new book shows that the Northern Dene people of Alaska and Canada have known far more about the stars than an earlier generation of scientists were willing to acknowledge.

A Traveler Waits in the Stars for Those Willing to Learn How to Look

The genetic testing company 23andMe has been fined more than £2.3m for failing to protect the personal information of more than 150,000 UK residents after a large-scale cyberattack in 2023.

Family trees, health reports, names and postcodes were among the sensitive data hacked from the California-based company. It only confirmed the breach months after the infiltration started and once an employee saw the stolen data advertised for sale on the social media platform Reddit, according to the UK Information Commissioner’s Office – which levied the fine.

The information commissioner, John Edwards, called the months-long incident across the summer of 2023 a “profoundly damaging breach”. The compromise of UK data was just a fraction of the wider losses, with the data of 7 million people affected.

23andMe charges users £89 to have their DNA screened using a saliva-based kit, allowing them to discover where their distant ancestors came from in terms of their ethnicity and location. But many customers asked for their DNA data to be deleted from the company’s archives after the hack and it filed for bankruptcy protection in the US in March.

The fine came as a $305m bid to buy the company led by its former chief executive, Anne Wojcicki, looked poised to retake control of the company in a bankruptcy auction.

Edwards said the data breach “exposed sensitive personal information, family histories and even health conditions of thousands of people in the UK”.

“As one of those impacted told us: once this information is out there, it cannot be changed or reissued like a password or credit card number,” he said.

23andMe failed to take basic steps to protect the information and their security systems were inadequate, the UK data protection regulator found. The breaches included failing to install tougher user authentication.

The hacker exploited a common weakness caused by users reusing passwords that had already been stolen in other unrelated data breaches. Hackers then used automated tools to try these passwords in a tactic called “credential stuffing”.

“The warning signs were there, and the company was slow to respond,” said Edwards, who carried out the investigation jointly with the privacy commissioner of Canada. “This left people’s most sensitive data vulnerable to exploitation and harm.”

skip past newsletter promotion Sign up to First Edition Free daily newsletter Our morning email breaks down the key stories of the day, telling you what’s happening and why it matters Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

A spokesperson for the company said 23andMe had since implemented multiple steps to increase security to protect individual accounts and information. They said that as part of the deal to acquire 23andMe, Wojcicki’s non-profit, the TTAM Research Institute, has made “binding commitments to enhance protections for customer data and privacy, including allowing individuals to delete their account and opt out of research at any time” and “agreeing not to sell or transfer genetic data under a subsequent bankruptcy or change of control”, and offering customers two years of free identity theft monitoring.

The fine is among several multimillion pound punishments meted out by the ICO in recent years for failure to protect data from hacks and ransomware attacks. In 2022, it fined the construction company Interserve £4.4m when staff data was compromised, including contact details, bank accounts, sexual orientation and health.

In March this year it fined an NHS IT supplier, Advanced Computer Software Group, nearly £3.1m for security failings that put the personal information of nearly 80,000 people at risk.

Information stolen from US company included details of 150,000 British residents including family trees

DNA testing firm 23andMe fined £2.3m by UK regulator for 2023 data hack

The US Department of Defense on Monday awarded OpenAI a $200m contract to put generative artificial intelligence (AI) to work for the US military.

The San Francisco-based company will “develop prototype frontier AI capabilities to address critical national security challenges in both warfighting and enterprise domains”, according to the defense department’s posting of awarded contracts.

The program with the defense department is the first partnership under the startup’s initiative to put AI to work in governments, according to OpenAI. The company plans to show how cutting-edge AI can vastly improve administrative operations such as how service members get healthcare and also cyber defenses, according to a blog post.

The startup claims that all use of AI for the military will be consistent with OpenAI usage guidelines, which are determined by OpenAI itself.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Big tech companies are increasingly pitching their tools to the US military, among them Meta and, more predictably, Palantir, the AI defense company founded by Peter Thiel, the conservative tech billionaire who has played a major role in Silicon Valley’s rightward shift.

OpenAI and defense tech startup Anduril Industries late last year announced a partnership to develop and deploy AI solutions “for security missions”. The alliance brings together OpenAI models and Anduril’s military tech platform to ramp up defenses against aerial drones and other “unmanned aircraft systems”, according to the companies.

“OpenAI builds AI to benefit as many people as possible, and supports US-led efforts to ensure the technology upholds democratic values,” OpenAI’s chief executive, Sam Altman, said at the time.

Program with the defense department is first under the startup’s initiative to put AI to work in governments

OpenAI wins $200m contract with US military for ‘warfighting’

Gary Marcus is right to point out – as many of us have for years – that just scaling up compute size is not going to solve the problems of generative artificial intelligence (When billion-dollar AIs break down over puzzles a child can do, it’s time to rethink the hype, 10 June). But he doesn’t address the real reason why a child of seven can solve the Tower of Hanoi puzzle that broke the computers: we’re embodied animals and we live in the world.

All living things are born to explore, and we do so with all our senses, from birth. That gives us a model of the world and everything in it. We can infer general truths from a few instances, which no computer can do.

A simple example: to teach a large language model “cat”, you have to show it tens of thousands of individual images of cats – being the way they are, they may be up a tree, in a box, or hiding in a roll of carpet. And even then, if it comes upon a cat playing with a bath plug, it may fail to recognise it as a cat.

A human child can be shown two or three cats, and from interacting with them, it will recognise any cat as a cat, for life.

Apart from anything else, this embodied, evolved intelligence makes us incredibly energy-efficient compared with a computer. The computers that drive an autonomous car use anything upwards of a kilowatt of energy, while a human driver runs on twentysomething watts of renewable power – and we don’t need an extra bacon sandwich to remember a new route.

At a time of climate emergency, the vast energy demands of this industry might perhaps lead us to recognise, and value, the extraordinary economy, versatility, plasticity, ingenuity and creativity of human intelligence – qualities that we all have simply by virtue of being alive.

Sheila Hayman

Advisory board member, Minderoo Centre for Technology & Democracy, Cambridge University

It comes as no surprise to me that Apple researchers have found “fundamental limitations” in cutting-edge artificial intelligence models (Advanced AI suffers ‘complete accuracy collapse’ in face of complex problems, study finds, 9 June). AI in the form of large reasoning models or large language models (LLMs) are far from being able to “reason”. This can be simply tested by asking ChatGPT or similar: “If 9 plus 10 is 18 what is 18 less 10?” The response today was 8. Other times, I’ve found that it provided no definitive answer.

This highlights that AI does not reason – currently, it is a combination of brute force and logic routines to essentially reduce the brute force approach. A term that should be given more publicity is ANI – artificial narrow intelligence, which describes systems like ChatGPT that are excellent at summarising pertinent information and rewording sentences, but are far from being able to reason.

But note, the more times that LLMs are asked similar questions, the more likely it will provide a more reasonable response. Again, though, this is not reasoning, it is model training.

Graham Taylor

Mona Vale, New South Wales, Australia

Letters: Sheila Hayman says artificial intelligence can’t match the qualities that we have just by the virtue of being alive, and Graham Taylor shows how AI can summarise but not reason

How AI pales in the face of human intelligence and ingenuity

I commend your reporting of the AI scandal in UK universities (Revealed: Thousands of UK university students caught cheating using AI, 15 June), but “tip of the iceberg” is an understatement. While freedom of information requests inform about the universities that are catching AI cheating, the universities that are not doing so are the real problem.

In 2023, a widely used assessment platform, Turnitin, released an AI indicator, reporting high reliability from huge-sample tests. However, many universities opted out of this indicator, without testing it. Noise about high “false positives” circulated, but independent research has debunked these concerns (Weber-Wulff et al 2023; Walters 2023; Perkins et al, 2024).

The real motivation may be that institutions relying on high-fee-paying international cohorts would rather not know; the motto is “see no cheating, hear no cheating, lose no revenue”. The political economy of higher education is driving a scandal of unreliable degree-awarding and the deskilling of graduates on a mass scale. Institutions that are biting the bullet, like mine, will struggle with the costs of running rigorous assessments, but know the costs of not doing so will be far greater.

If our pilots couldn’t fly planes themselves or our surgeons didn’t know our arses from our elbows, we’d be worried – but we surely want our lawyers, teachers, engineers, nurses, accountants, social workers etc to have real knowledge and skills too.

A sector sea change is under way, with some institutions publicly adopting proper exams (maligned as old-fashioned, rote-learning, unrealistic etc) that test what students can actually do themselves. Institutions that are resistant to ripping off the plaster of convenient yet compromised assessments will, I’ll wager, have to some day explain themselves to a public inquiry.

Dr Craig Reeves

Birkbeck, University of London

Letter: Dr Craig Reeves says some institutions don’t seem interested in dealing with the problem of AI use by students and are resisting in-person assessments

Universities face a reckoning on ChatGPT cheats

I doubt that government officials consulted their AI tool, Humphrey, on what it should be called (UK government rollout of Humphrey AI tool raises fears about reliance on big tech, 15 June). It could have advised that in the 1970s the name was used for a milk marketing campaign: “Watch out, there’s a Humphrey about.” That line will now have a whole new meaning. Having spent the last few weeks voting in the Lords to try, in vain, to achieve protections for the creative industries from AI abuse, that meaning might be prophetic. On a personal level, my husband is angry that his name is being stolen again.

Sue Miller

Liberal Democrat, House of Lords

Your report (The ‘death of creativity’? AI job fears stalk advertising industry, 9 June) quotes the chief executive of a big ad agency saying “AI will disintermediate a large number of jobs”. I asked an AI tool to put this into plain English. Ten minutes later, it’s still thinking about it. Perhaps absurd management euphemisms is one area where the human brain can still cut through the crap more effectively than AI?

Neil Cole

Hemel Hempstead, Hertfordshire

Heather Stewart says large language models, while destroying creative jobs and careers, “remain prone to casually making things up” (Policymakers who think AI can help rescue flagging UK economy should take heed, 15 June). Could her article spur MPs into taking meaningful legislative action to protect their own positions?

Jonathan J Ross

Sheffield

Re Heather Stewart’s article, I’m not a specialist, but has anyone tried unplugging AI, going for a walk and then going back to check what’s happened? If the door’s locked, we’ve had it.

Gavin Greenwood

Brighton

To check one AI’s “hallucination” flaws, can we not use another AI to check what the first AI has produced, and possibly made up?

Alan Worsley

Hull

As the English language still seems to lack a verb meaning “to generate something via AI”, might I suggest “to bot”? This would provide a quick means of asking whether someone really did the work themselves or just botted it.

Edmund Doragh

Schwäbisch Gmünd, Germany

Watch out, hallucinating Humphrey’s about in Whitehall

In your report (High court tells UK lawyers to stop misuse of AI after fake case-law citations, 6 June), you quote Dame Victoria Sharp’s call that we, the Bar Council, and our solicitor colleagues at the Law Society address this matter urgently.

We couldn’t agree more. This high court judgment emphasises the dangers of the misuse by lawyers of artificial intelligence, particularly large language models, and in particular its serious implications for the administration of justice and public confidence in the justice system.

The public is entitled to expect from legal professionals the highest standards of integrity and competence in appropriate understanding and use of new technologies, as well as in all other respects.

The Bar Council has already issued guidance on the opportunities and risks surrounding the use of generative AI, which is quoted by the court, and is in the process of setting up a joint working group with the Bar Standards Board to identify how best we can support barristers to uphold those standards with appropriate further training and supervision.

Lawrence Akka KC

Chair, Bar Council IT panel

Letter: Responding to an article on fake case-law citations in court, Lawrence Akka KC writes that the public is entitled to the highest standards of integrity

Bar Council is wise to the risk of AI misuse

Elon Musk’s X Corp filed a lawsuit on Tuesday against the state of New York, arguing a recently passed law compelling large social media companies to divulge how they address hate speech is unconstitutional.

The complaint alleges that bill S895B, known as the Stop Hiding Hate Act, violates free speech rights under the first amendment. The act, which the governor, Kathy Hochul, signed into law last December, requires companies to publish their terms of service and submit reports detailing the steps they take to moderate extremism, foreign influence, disinformation, hate speech and other forms of harmful content.

Musk’s lawyers argue that the law, which goes into effect this week, would require X to submit “highly sensitive information” and compel non-commercial speech, which is subject to greater first amendment protections. The complaint also opposes the possible penalty of $15,000 per violation per day for failing to comply with the law.

The bill’s authors issued a statement on Tuesday rejecting Musk’s suit as an attempt to “use the First Amendment as a shield against providing New Yorkers with much needed transparency” and argued that the act does not infringe upon the free speech rights of social media companies.

X successfully sued last year to block a similar law in California, which also required social media companies to disclose to the government how they define hate speech and extremism. Since Musk bought Twitter in 2022, he has heavily downgraded content moderation efforts and fought against attempts at regulating the platform.

The Stop Hiding Hate Act was written by Brad Hoylman-Sigal, a New York state senator and Grace Lee, a state assemblymember, in cooperation with the Anti-Defamation League, with the intent of providing great transparency and accountability over harmful online speech. Lee and Hoylman-Sigal rejected a request last year from X to discuss and amend the bill, according to the state senator, on the grounds that they believed the company was not acting in good faith and only seeking to weaken the bill’s requirements.

“Now more than ever, with the rise in political violence and threats emanating from the spread of hate speech and disinformation by President Trump and Elon Musk, New Yorkers deserve to know what social media companies like X are doing (or not doing) to stop the spread of hatred and misinformation on their platform,” Hoylman-Sigal and Lee said Tuesday.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Although Musk has described himself as a “free speech absolutist”, his platform X has targeted journalists and media outlets that post critical information about him. The platform suspended the accounts of several journalists in recent years that have reported on Musk and throttled links to news sites he has criticized. Musk also sued a non-profit that catalogued racist and extremist content on X, in a case that a US district judge threw out as an attempt to punish freedom of speech.

Suit alleges Stop Hiding Hate Act, which compels social media firms to disclose actions against hate speech, violates free speech

Elon Musk’s X sues New York over hate speech and disinformation law

The Metropolitan police’s recognition of the value in “some sort of framework or statutory guidance” for live facial recognition is welcome (Live facial recognition cameras may become ‘commonplace’ as police use soars, 24 May). However, it is not just police use of this technology that needs a clear legal framework.

Despite the scale and speed of its expansion, there is still no specific law providing a basis for live facial recognition or other emerging biometric technologies, whether these are used in the public or private sector.

Biometric surveillance is expanding rapidly, not just in policing but across society: in train stations, schools and supermarkets. Newer biometric systems go further, claiming to infer people’s emotional states, raising serious concerns about their accuracy, ethics and legality.

In 2020, the court of appeal – in the UK’s only judgment on live facial recognition to date – found a deployment by South Wales police was unlawful. It identified deficiencies in the legal framework and set out minimum standards for lawful use. Since then, a patchwork of voluntary guidance has emerged.

New research from the Ada Lovelace Institute has found this patchwork is inadequate in practice, creating legal uncertainty, putting fundamental rights at risk and undermining public trust.

Crucially, we find that non-police uses, such as those in the private sector or involving inference, are subject to even fewer safeguards and so stand on far shakier legal ground. Governance is simply not keeping up with technological adoption and advancement.

Policymakers must act. We urgently need new legislation to establish a comprehensive framework covering all forms of biometric surveillance and inference – not just police use – and an independent regulator to oversee and enforce it.

Michael Birtwistle

Associate director, the Ada Lovelace Institute

Letters: The rapid expansion of biometric surveillance raises serious concerns about accuracy and ethics, but oversight is patchy and voluntary, says Michael Birtwistle

Facial recognition technology needs stricter regulation

Reddit celebrates its 20th birthday at the end of the month. With 17 million daily viewers, the online community forum has brought various issues to people’s attention, from the timely and topical to the bizarre.

We’d like to hear about your best Reddit moment. Perhaps you’ve been able to share a personal experience and felt you found your tribe discussing it with others. Or maybe you’ve had a complex issue explained to you like a five-year-old, or just found yourself laughing along with a viral moment with millions of others.

Share your experience You can tell us about your best Reddit moment using this form. Please share your story if you are 18 or over, anonymously if you wish. For more information please see our terms of service and privacy policy Tell us here Your responses, which can be anonymous, are secure as the form is encrypted and only the Guardian has access to your contributions. We will only use the data you provide us for the purpose of the feature and we will delete any personal data when we no longer require it for this purpose. For true anonymity please use our SecureDrop service instead. Name Where do you live? Tell us a bit about yourself (e.g. age and what you do for a living) Optional Tell us about your best Reddit moment Please include as much detail as possible. If you are happy to, please upload a photo of yourself here Optional Please note, the maximum file size is 5.7 MB . Choose file Can we publish your response? Yes, entirely Yes, but contact me first Yes, but please keep me anonymous No, this is information only Phone number Optional Your contact details are helpful so we can contact you for more information. They will only be seen by the Guardian. Email address Your contact details are helpful so we can contact you for more information. They will only be seen by the Guardian. You can add more information here Optional If you include other people's names please ask them first. Would you be interested in speaking to our audio and/or video teams? Audio only Video only Audio and video No, I'm not interested By submitting your response, you are agreeing to share your details with us for this feature. Submit

If you’re having trouble using the form click here. Read terms of service here and privacy policy here.

As Reddit celebrates its 20th birthday at the end of the month, we’d like to hear about your favourite moments on the online community forum

Tell us about your best Reddit moment

Bats carry an assortment of viruses, including infectious diseases that have jumped into humans in a process known as zoonotic spillover. In some cases, the aerial mammals have infected people directly. But in other cases, scientists believe, bats have passed deadly viruses onto other animal species, which subsequently infected humans.

A remarkable new video from Uganda provides direct visual evidence of many different wild animals interacting with wild bats, revealing just how many opportunities there are for potentially deadly bat viruses to jump into new species. It was published along with a paper posted online on Monday but not yet peer reviewed.

On Feb. 17, Bosco Atukwatse, a young Ugandan wildlife biologist working with the Kyambura Lion Project, set up solar-powered camera traps near the mouth of Python Cave in the Maramagambo Forest, in Queen Elizabeth National Park in Uganda. He was collecting data about leopards and spotted hyenas in the forest.

Through a mixture of what he called “curiosity and luck,” he filmed far more than leopards. Hundreds of nights of footage revealed a steady procession of 13 additional predator species, among them large-spotted genets, African civets, African fish eagles, African rock pythons, L’Hoest’s monkeys and baboons. Python Cave is home to as many as 50,000 Egyptian fruit bats, and the predators emerged from the cave with a winged snack, which they either hunted or scavenged, in their mouths.

Video from a national park in Uganda depicted a parade of predatory species feeding on and dispersing fruit bats that are known natural reservoirs of infectious diseases.

Bat Cave Footage Offers Clues to How Viruses Leap Between Species

Ms. Morrow, 24, from Powder Springs, Ga., is one of hundreds of people who wrote to their hometown newspapers as part of a national campaign spearheaded this spring by graduate students and scientists who are just starting their careers.

They wanted to draw attention to the Trump administration’s research funding cuts that are scuttling grants, shrinking science labs and stopping postdoctoral studies. Administration officials have pointed to the importance of cost-saving and attributed many of the budget cuts to changing scientific priorities. The White House has moved to cancel research in specific areas, like transgender health and climate science, and described some research efforts as wasteful spending.

For Emma Scales, a graduate student who studies plant pathology and plant-microbe biology at Cornell, the slashed budgets evoked an unpleasant realization. “Oh, my God,” she remembered thinking. “Nobody knows what we do.”

This, she added, was partly the fault of scientists — a communication failure that she felt determined to address. “Here I am,” she said. “Let me tell you. I want to show you what it is that I do, and why it’s important.” (Her research on fungal-bacterial endosymbiosis could one day help to protect humans and harvests from dangerous pathogens.)

Ms. Scales became one of the organizers of a project called the McClintock Letters, with many to be published on the June 16 birthday of the geneticist Barbara McClintock, who in 1983 became the first woman to be awarded an unshared Nobel Prize for Physiology or Medicine.

Hundreds of graduate students are writing to their hometown newspapers to defend their research, as the Trump administration drastically reduces science funding.

Graduate Students Push Back Against Science Funding Cuts

The proceeds of the sale, he said, “are critical and necessary to our ongoing sustainability, and exclusively reserved and used to maintain and protect our collections, educate visitors and further expand our collections.”

“This juvenile ceratosaur is a truly remarkable specimen,” said Cassandra Hatton, vice chairman and global head of science and natural history at Sothebys. She noted that because the Museum of Ancient Life lacks paleontological certification as a public repository, the specimen has never been fully described or studied. She pointed to the sale of “Sue” the Tyrannosaurus rex to the Field Museum in Chicago and the current four-year loan of “Apex” the stegosaurus to the American Museum of Natural History as evidence of the auction house’s “track record in helping to get specimens such as this into the public trust.”

Mr. Sisson and his team rebuilt the skeleton, used 3-D printing and sculpted elements to fill in the missing bones and mounted the genuine fossils in jewelry-grade metal brackets, allowing the bones to be removed individually and, he said, preserving them for research purposes.

As with “Apex,” the specimen will be offered with supplementary records and documentation of its original excavation by Western Paleontological Laboratories, and Mr. Sisson’s reconstruction. That paperwork helps maintain its scientific integrity, Ms. Hatton said.

It also increases the fossil’s value.

Andre LuJan, president of the Association of Applied Paleontology, a trade group that represents preparators, fabricators and many commercial paleontologists, said that raising the threshold for documentation on dinosaurs brought to market had had a positive effect by preventing some disputed material from going to auction.

The 150-million-year-old specimen is valued at up to $6 million by Sotheby’s. Some paleontologists worry this auction and earlier ones are driving fossil market speculators.

Sotheby’s to Auction a Ceratosaurus, With Millions and More on the Line

The city’s most famous sons may have sung that money can’t buy you love, but that was before bitcoin existed.

Liverpool has emerged as the crypto capital of the UK, according to a study looking at the online habits of people across the country.

The survey, conducted by telecommunications company Openreach, found that 13% of respondents from Liverpool regularly invest in cryptocurrency and check stocks, more than anywhere else in Britain.

Different cities across the UK proved to be hotspots for various activities. London seems to be the online dating capital of Britain, with 24% of respondents saying they engage with dating apps on at least three days a week.

This contrasts with the country in general, with the study finding that only 4% of Britons spend any time on dating apps.

According to the study, the average British person claims to spend three-and-a-half hours a day online, though 20% of those asked admitted to spending above five hours of their day on the internet.

Popular times to be online varied, but 64% of respondents said they spend time online between the hours of 11pm and 6am, with 19% of them saying this is the time they visit YouTube.

In the north, analysis tells a tale of three cities. The people of Manchester used Instagram more than anywhere else, with 27% of people in the city using the platform regularly.

Sheffield, meanwhile, is home to both the most frequent TikTok users and music streamers in the country (with figures of 32% and 30% respectively). Sheffield is also the city where households spend the most time online, with 32% spending more than five hours online per day – in contrast to the 11% of Brighton citizens who do the same.

The people of Leeds seem to favour yesterday’s social media site of choice, with 43% of the city’s residents saying they spend a lot of time using Facebook.

The study was commissioned by Openreach to coincide with the first installation of broadband in a UK home, which took place in April 2000 in Basildon, Essex.

The study also found that many respondents disliked the way they use the internet, with 43% feeling that they wasted time online, 37% concerned by the hours spent “doom-scrolling” and 33% saying they would feel more relaxed if they spent less time online.

Katie Milligan, deputy CEO of Openreach, said: “It’s fascinating to see how different parts of the UK are embracing the online world and adapting to it in unique ways.

“At the same time, it’s encouraging that many recognise the importance of taking time away from devices and digital connectivity.”

Research reveals 13% of residents regularly invest in cryptocurrency and check stocks, more than all other cities

Liverpool is crypto capital of UK, survey finds

The Guardian’s Latin America correspondent, Tom Phillips, recalls the moment that he and others on the search team found Dom and Bruno’s belongings in a hidden area of flooded forest. The team finally discover what has happened to the men

Episode four: the ambush

Makers of air fryers, smart speakers, fertility trackers and smart TVs have been told to respect people’s rights to privacy by the UK Information Commissioner’s Office (ICO).

People have reported feeling powerless to control how data is gathered, used and shared in their own homes and on their bodies.

After reports of air fryers designed to listen in to their surroundings and public concerns that digitised devices collect an excessive amount of personal information, the data protection regulator has issued its first guidance on how people’s personal information should be handled.

It is demanding that manufacturers and data handlers ensure data security, are transparent with consumers and ensure the regular deletion of collected information.

Stephen Almond, the executive director for regulatory risk at the ICO, said: “Smart products know a lot about us: who we live with, what music we like, what medication we are taking and much more.

“They are designed to make our lives easier, but that doesn’t mean they should be collecting an excessive amount of information … we shouldn’t have to choose between enjoying the benefits of smart products and our own privacy.

“We all rightly have a greater expectation of privacy in our own homes, so we must be able to trust smart products are respecting our privacy, using our personal information responsibly and only in ways we would expect.”

The new guidance cites a wide range of devices that are broadly known as part of the “internet of things”, which collect data that needs to be carefully handled. These include smart fertility trackers that record the dates of their users’ periods and body temperature, sends it back to the manufacturer’s servers and makes an inference about fertile days based on this information.

Smart speakers that listen in not only to their owner but also to other members of their family and visitors to their home should be designed so users can configure product settings to minimise the personal information they collect.

skip past newsletter promotion Sign up to First Edition Free daily newsletter Our morning email breaks down the key stories of the day, telling you what’s happening and why it matters Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The regulator said manufacturers needed to be transparent with people about how their personal information was being used, only collect necessary information and make it easy for people to delete their data from the product.

The ICO told manufacturers “we are ready to take action if necessary to protect people from harm”.

Information Commissioner’s Office takes action as people report feeling powerless over data gathering at home

Makers of air fryers and smart speakers told to respect users’ right to privacy

23andMe’s former CEO is set to regain control of the genetic testing company after a $305m bid from a non-profit she controls topped a pharmaceutical company’s offer for it in a bankruptcy auction.

Last month, Regeneron Pharmaceuticals agreed to buy the firm for $256m, topping a $146m bid from Anne Wojcicki and the non-profit TTAM Research Institute. The larger offer prompted Wojcicki to raise her own with the backing of a Fortune 500 company, according to the former executive. The deal is expected to close in the coming weeks after a court hearing currently scheduled for 17 June, the company said on Friday.

Wojcicki had made multiple bids to take the company private while still CEO. The company’s board rejected her each time, and all of its independent directors eventually resigned in response to her takeover attempts.

Once a trailblazer in ancestry DNA testing, 23andMe filed for bankruptcy in March, seeking to sell its business at auction after a decline in demand and a 2023 data breach that exposed sensitive genetic and personal information of millions of customers.

23andMe has lost a major chunk of customers since declaring bankruptcy, with the threat of a sale of users’ most sensitive information – the company sequences a person’s entire genome – to an unknown buyer looming. The company has said that about 15% of its existing customers have requested the closure of their accounts in response to its bankruptcy and planned sale. Experts have advised customers to ask the company delete their DNA data for privacy protection. TTAM said on Friday it would uphold 23andMe’s existing privacy policies and comply with all applicable data protection laws. Earlier this week, New York and more than two dozen other US states sued 23andMe to challenge the sale of its customers’ private information.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Regeneron had said it was willing to make a new bid, but wanted a $10m breakup fee if Wojcicki’s bid was ultimately accepted.

Anne Wojcicki made $305m bid for firm, which has lost customers since declaring bankruptcy, with backing of Fortune 500 company

23andMe’s founder wins bid to regain control of bankrupt DNA testing firm

Mark Zuckerberg announced in April that the company would make huge capital expenditures in the coming year to keep up in the race to develop cutting-edge artificial intelligence. He made good on that promise last week with a $15bn “AI superintelligence” team that would feature reported nine-figure salaries and a 49% investment in Scale AI. Meta also hired Scale’s 28-year-old founder, Alexandr Wang, a former roommate of OpenAI’s Sam Altman.

Before Meta’s investment, Scale counted most of the major players in AI among its clients, and some of them were less than thrilled with the development. Bloomberg puts it succinctly: Scale AI’s Wang Brings to Meta Knowledge of What Everyone Else is Doing. Google, Scale’s largest customer, got scared. The tech giant told the startup that their working relationship would end in response to the deal, Reuters reported on Friday.

My colleague Robert Booth has more:

One Silicon Valley analyst described the announcement as the action of “a wartime CEO”.

Superintelligence is described as a type of AI that can perform better than humans at all tasks. Currently AI cannot reach the same level as humans in all tasks, a state known as artificial general intelligence (AGI).

Meta’s attempt to leapfrog the current state of progress and target superintelligence is seen by observers as an attempt by the company to regain the initiative over AI after significant advances by competitors including Sam Altman’s OpenAI and Google and after Meta’s huge investment in the concept of the Metaverse flopped.

The financial magnitude of the move may seem like taking command of the AI race, but Meta is playing catchup. Its most recent AI models are less capable than those of rivals. In April, the company published a model designed with features tailored to gaming a popular benchmark, a desperate move. With its mammoth investment and the formation of a new team, particularly one helmed by a key industry player, the tech giant is buying its way in.

Read more about Meta’s superintelligence team here.

The week in AI news

Wikipedia, vital to AI, takes pains to highlight human contributions

View image in fullscreen Photograph: Blake Montgomery

Last week, Wikipedia halted a test of a new AI-powered article summarizer in response to a backlash from its editors, the volunteers who contribute new information to the online encyclopedia’s articles and evaluate the quality of others’ additions. The site had introduced summaries of articles, generated by artificial intelligence, that would appear at the top of those pages above the introduction to the subject, written by humans. The test ran for two weeks for about 10% users of the mobile version of Wikipedia who opted into it.

The feedback was, in the words of Wikipedia’s nonprofit parent, “strongly negative”. One editor wrote in the public forum: “A truly ghastly idea.” Another: “Keep AI out of Wikipedia.” And: “I am going to join the chorus of editors saying that this is a uniquely bad idea.” Some argued in favor of the introduction of summaries by AI, making the point that technical articles on Wikipedia prioritize depth and complexity over accessibility. Most, though, argued that any hallucinations and mistakes would degrade Wikipedia’s reputation of reliability.

The general sentiment among editors read thusly: “Yuck”.

“We’re trying to figure out the balance between AI and Wikipedia. There are AI integrations that do happen on Wikipedia but are very minimal. Little things that make it easier for editors to do their work or readers to get the content they need,” said Pacita Rudder, executive director of Wikimedia New York City, a local chapter affiliated with the foundation. “‘Where’s the line between human output and robotic output?’ is a debate that’s constantly happening within the Wikipedia community. There are some people who are for it, some against it, but as with anything with Wikipedia, we have to come to a consensus.”

Contrast Wikipedia’s deferential response with Reddit’s. The social network began charging for access to its application programming interface in 2023. Reddit’s library of posts is extremely valuable to AI companies, which need enormous collections of text written by contemporary human beings to train large language models like ChatGPT. Reddit execs intended to extract money from AI companies using the company’s library. They succeeded. They also rendered many users’ experiences worse by crippling popular extensions. In response, users, particularly forums’ volunteer moderators, revolted en masse, blacking out forums with tens of millions of members and demanding reversion. Executives refused. They chose dollars from AI companies over their users’ concerns, the opposite of Wikipedia. A coda to the story, though: two years later, Reddit largely functions as before.

Several days after Wikipedia rolled back its AI test, the Wikimedia Foundation hosted an in-person “edit-a-thon” at the United Nations, an event dedicated to creating and updating pages about the history of the international body, complete with a work list. The conclave of information custodians – a 50/50 balance of new and experienced editors, according to Rudder – was part of the UN’s Open Source Week and occurred alongside an open source hackathon. It was in-person only.

The edit-a-thon stands in contrast to the stereotypical tableau of curating Wikipedia articles: a single person online, communicating sterilely with others via a glaring white screen. At the event, though, editors convened in groups of five around a single laptop to converse excitedly and collaborate in as many languages throughout a large, buzzing auditorium. Italian and Sri Lankan envoys lionized the work of the editors in speeches, as well as the open source community at large. An executive from the Wikimedia Foundation thanked them for their work, as did the UN under-secretary-general for digital and emerging technologies. It was an energetic scene.

Asked why she felt it was important to host an in-person-only event, Rudder said: “It’s community, right? There’s this misconception that a lot of Wikipedia editors are alone in their basements, typing away, but that’s not the case,” said Rudder. Wikimedia NYC hosted the event at the UN. Rudder said the chapter’s role is to bring editors and members of the online community together. She said she felt most energized by meeting new editors.

As with Reddit, Wikipedia’s huge text corpus is part of the foundational training data for most, if not all, large language models. And like the white alien company, Wikipedia charges for smooth access to it, though less than Reddit, Rudder said. The organization launched Wikimedia Enterprise in 2021 with an eye towards corporate customers. Reddit is a for-profit business, whereas the Wikimedia Foundation is a non-profit. Wikipedia’s data is free, but using it easily may cost you.

Unlike Reddit, Wikipedia is emphasizing the human labor necessary to maintain its site and the content there by rolling back features in response to community outcry and convening editors in person.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The tech angle on the Los Angeles anti-deportation protests

Apple’s walled garden turns from Eden to a pen

View image in fullscreen Photograph: Apple Corps Ltd.

A week ago, Apple hosted its annual developers’ conference, WWDC. The biggest announcement was a redesign of the iPhone Operating System (iOS) nicknamed “liquid glass”. Virtual buttons on iPhone screens will appear transparent, with the images behind them passing through with a slight warp like a magnifying glass. The aesthetic recalls the colorful Macs of old, housed in translucent, whimsical plastic.

The refreshed design is pretty! It is also boring. Machines are fabricating entire movies and predicting hurricanes. Who cares what color my phone’s controls are?

There were other updates, of course. Apple is adding live translation to FaceTime calls, for instance. I can imagine many calls with distant family and friends that will run smoother with that feature. How nice.

On the whole, though, the updates failed to inspire. For many years, updates to iPhones have seemed far more incremental than, for example, the addition of speedy 3G connectivity in 2008 with the second generation. Apple’s updates have failed to keep pace with the freewheeling Android ecosystem. I met a venture capitalist in San Francisco in 2019 who did not know what version of the iPhone he was holding in his hand as we spoke.

Apple showcased little in the way of new AI features, unlike its peers. Google and Samsung are integrating AI into their mobile experiences at a breakneck pace. Case in point: live translation is a feature that has been part of the Android operating system for several years. Google’s Pixel Bud headphones can translate the speech of a person talking to you, which Airpods cannot do. Google seems to release a fascinating, if sometimes bizarre, AI feature every week. As recently as Friday, the Verge reported: Google says it is testing Audio Overviews for certain Search queries, available for Labs users in English in the US.

After the demonstration, Gizmodo argued: Apple Knows AI Isn’t What People Really Want, but It Can’t Say That. Even if that is true, though – and ChatGPT’s 500m monthly users might disagree – the announcements showed little in the way of flair. If Apple wants users to stay within the iPhone’s ecosystem, it needs to offer them something equally appealing and enthralling or even more so.

Google pushed a new, updated version of the operating system and last week. Samsung teased new features of the next version of its folding phone, per Android Central. Did you hear about them? It is axiomatic in tech journalism that stories about Apple and iPhones will always garner more clicks than stories about Android phones, though a far higher percentage of the world’s phones use Android. I have many guesses as to why that is, but it is a topic for another day. I only mention it because many fewer people spent the week discussing the Android updates as did Apple’s Liquid Glass interface. Apple’s change amounts to only a cosmetic upgrade – and it’s debatable whether it will be better at all – but it inspired many more jokes.

Samsung’s seventh version of its folding phone, while novel, begs the question: as I complain about new iPhones’ lack of thrills, what do I want my phone to do? I do not need a folding phone. I had one, a Motorola Razr, and it did facilitate vital communication; however, that did not arise from the form factor.

Google’s new AI features pose the same question. There may not be consensus on what the majority of users want from an AI-enhanced phone, but if you are not trying something new with technology, you are in decline. It doesn’t feel like Apple is offering surprising features that I would not have thought of. I may not need an audio summary feature for my Google search results, but I am intrigued by it. Also, what a gift to the visually impaired! Siri, by contrast, still struggles to control Spotify and botches notification summaries.

I have used an iPhone as my primary device for about 15 years now. This inflection point would be the moment to switch to Android to try out its shiny toybox of new features, but I am enclosed in Apple’s ecosystem of hardware and apps, a walled garden that is starting to feel less like a bountiful Eden and more like a pen.

The wider TechScape

The magnitude of Meta’s investment in Scale may seem like command of the AI race, but the company’s playing catchup

Meta sacrifices a heap of money at the altar of AI

The government’s artificial intelligence (AI) tool known as Humphrey is based on models from OpenAI, Anthropic and Google, it can be revealed, raising questions about Whitehall’s increasing reliance on big tech.

Ministers have staked the future of civil service reform on rolling out AI across the public sector to improve efficiency, with all officials in England and Wales to receive training in the toolkit.

However, it is understood the government does not have overarching commercial agreements with the big tech companies on AI and uses a pay-as-you-go model through its existing cloud contracts, allowing it to swap through tools as they improve and become competitive.

Critics are concerned about the speed and scale of embedding AI from big tech into the heart of government, especially when there is huge public debate about the technology’s use of copyrighted material.

Ministers have been locked in a battle with critics in the House of Lords over whether AI is unfairly being trained on creative material without credit of compensation. Its data bill allowing copyrighted material to be used unless the rights holder opts out passed its final stage this week in a defeat for those fighting for further protections.

The issue has caused a fierce backlash from the creative sector, with artists including Elton John, Tom Stoppard, Paul McCartney and Kate Bush throwing their weight behind a campaign to protect copyrighted material.

A freedom of information request showed the government’s Consult, Lex and Parlex tools designed to analyse consultations and legislative changes use base models from Open AI’s GPT, while its Redbox tool, which helps civil servants with everyday tasks such as preparing briefs, uses Open AI GPT, Anthropic’s Claude and Google Gemini.

Ed Newton-Rex, the chief executive of Fairly Trained, who obtained the FoI and is campaigning against AI being trained on copyrighted material, said there was the potential for a conflict when the government was also thinking about how this sector should deal with copyright.

He said: “The government can’t effectively regulate these companies if it is simultaneously baking them into its inner workings as rapidly as possible. These AI models are built via the unpaid exploitation of creatives’ work.

“AI makes a ton of mistakes, so we should expect these mistakes to start showing up in the government’s work. AI is so well known for ‘hallucinating’ – that is, getting things wrong – that I think the government should be keeping transparent records of Humphrey’s mistakes, so that its continuing use can be periodically reevaluated.”

Shami Chakrabarti, the Labour peer and civil liberties campaigner, also urged caution and to be mindful of biases and inaccuracies such as those seen in the Horizon computer system that led to the miscarriage of justice for post office operators.

Whitehall sources said Humphrey tools all worked in different ways, but users could take different approaches to tackling “hallucinations”, or inaccuracy, and the government continually publishes evaluations about the accuracy of technology in trials. An AI playbook for government also sets out guidance to help officials make use of the technology quickly and offers advice on how to ensure people have control over decisions at the right stages.

The costs of using AI in government are expected to grow as Humphrey is further rolled out but officials say prices of AI per-use in the industry have trended downwards, as models become more efficient.

Whitehall sources said big projects such as the Scottish government’s use of AI to analyse consultation responses had cost less than £50 and saved many hours of work.

Using the government’s AI Minute software to take notes for a one-hour meeting costs less than 50p and its early data shows that it saves officials an hour of admin each time.

A spokesperson from the Department for Science, Innovation and Technology said: “AI has immense potential to make public services more efficient by completing basic admin tasks, allowing experts to focus on the important work they are hired to deliver.

“Our use of this technology in no way limits our ability to regulate it, just as the NHS both procures medicines and robustly regulates them.

“Humphrey, our package of AI tools for civil servants, is built by AI experts in government – keeping costs low as we experiment with what works best.”

When the Guardian asked ChatGPT what base models were used for the Humphrey AI toolkit and if Open AI was involved, it replied that the information was not available.

At the time the tool was announced earlier this year, the government said its strategy for spending £23bn a year on technology contracts would be changed, boosting opportunities for smaller tech startups.

Critics question speed and scale of embedding AI into public sector when debate rages about use of copyrighted material

UK government rollout of Humphrey AI tool raises fears about reliance on big tech

Artificial intelligence seems to have touched every part of our lives. But can it help us manage our money? We put some common personal finance questions to the free version of ChatGPT, one of the most well-known AI chatbots, and asked for its help.

Then we gave the answers to some – human – experts and asked them what they thought.

Pensions

We asked: I am 35 years old and want to ensure I have a comfortable retirement. I earn about £35,000 a year and have a workplace pension, in which I have saved £20,000. My employer contributes 4% of salary. I would like to retire at 60 and want to ensure I have enough in my pension pot to not worry about money later in life. Do I need to save more?

ChatGPT said*: A comfortable single-person retirement requires roughly £43,000 a year of spending (in today’s money) while a moderate lifestyle might need £31,000 a year. Covering £43,000 a year over a 20- to 30-year retirement implies a very large pension pot of about £793,000 in addition to the state pension. A moderate £31,000 a year needs £478,000 in savings.

Your state pension will kick in at 67 and is about £12,000 a year, much less than the £43,000 needed for comfort and it does not arrive until seven years after you retire. This “pre-state pension gap” could well exceed £150,000. You must plan to fund all expenses for seven years on your own savings.

Even optimistic growth leaves you well below the £500,000 to £800,000 needed. You’ll need to save more – much higher than 9%. Vanguard [an investment adviser] suggests saving 12% to 15% of your pay a year (including employer contributions). Legal & General offers a rule-of-thumb savings percentage of half your age: at 35, that means 17.5% of your salary. Since your employer already gives 4%, you might target 11%. Don’t forget UK tax relief: your 20% basic-rate relief effectively makes every £80 you save cost you only £60 net.

Consider topping up with a lump sum or contributing to a stocks and shares Isa.

Read ChatGPT’s full answer here

View image in fullscreen ChatGPT’s responses were largely limited to general summaries and sometimes included inaccuracies. Photograph: Nicolas Economou/NurPhoto/REX/Shutterstock

Steve Webb, a partner at the pension consultants LCP, says: Using the Pensions and Lifetime Savings Association (PLSA) standard for a “comfortable” and moderate income seems a reasonable place to start.

For someone aged 35 today, the state pension age by law will be 68 (and could be higher by the time they get there) so the reference to 67 in the reply is incorrect.

The PLSA says you need £43,000 a year after tax, so that implies a pre-tax income of very roughly £50,000; you get £12,000 a year of this from state pension age, but for the first eight years of retirement (before the state pension kicks in) you would need this to come from your pension pot. This means you will need £635,000 on current calculations.

The figure of £793,000, therefore, looks over the odds, though there may be an assumption (not made explicit) that a quarter is always taken in tax-free cash. The person using AI needs to understand what assumption has been made around tax-free cash to clarify what is going on.

It is good the AI explains that in the first years of retirement you are relying solely on your private pension pot, a kind of basic “cashflow modelling” many advisers would do as standard.

The rules of thumb are a decent starting point, although the one about saving at a percentage which is half your age is for people who only start saving at that age.

One omission is that there is no reference to the possibility of an additional employer match. Some firms will pay more if you pay more, and that would make additional workplace saving much more attractive than the suggested alternative of an Isa. Another omission is that if you buy an annuity at 60, you may want to think about things such as inflation protection, and these would increase the size of pot you needed to target.

I don’t think it’s a bad reply aside from the error on state pension age. But you would want to go deeper in terms of understanding the assumptions made.

Mortgages

We asked: We have 20 years left on our mortgage and our current rate is due to run out in a few months. The loan to value is about 50%. What are the best rates on the market? Should we go for a two-year or five-year fix? Or should we stay put and hope that that rates come down?

ChatGPT said: The very cheapest two‑year fixes are about 3.7% to 3.8% and five‑year fixes are about 3.8% to 3.9% for borrowers with large deposits. With 50% equity you should qualify for the lowest-rate tier.

For example, Lloyds Bank has a two‑year fix at 3.72%. These often have arrangement fees (typically £999) so factor them into any calculation. For five‑year fixes, Lloyds offers 3.78%.

If you did nothing, your mortgage would revert to the lender’s standard variable rate, which are typically much higher – about 7-8%. An SVR can jump if rates rise, and lenders are not required to pass on cuts in full.

If you expect further rate cuts, a two‑year deal lets you refinance again sooner at (hopefully) lower rates. The downside is more frequent remortgaging: after two years, you’ll need another deal.

Five-year deals are likely to be a few tenths higher but give peace of mind for longer and protect you if market rates jump or stay elevated. If stability of repayments is your priority, a five‑year fix is often recommended.

Most analysts expect further modest cuts – economists at Barclays predicted the bank rate could fall to 3.5% by early 2026. That could push mortgage offers down further, but likely not to the ultra-low levels seen pre-2020.

Many experts lean toward taking a fixed rate now (especially to avoid the very high SVR) and revisiting options after a year or two.

Read ChatGPT’s full answer here

View image in fullscreen ChatGPT can be helpful as the starting point of broader research into mortgages. Photograph: Dominic Lipinski/PA

David Hollingworth of the broker L&C Mortgages says: This response is not even close to the level of advice that a qualified adviser would provide. What it does provide is a quick spin around some of the things that have been going on in the market, along with a general overview of elements the borrower may want to think about.

The AI response is a bit of a mishmash and veers from quite specific information, such as specifying the Lloyds rate, to more general observations. Aside from the fact that this deal is now defunct, it also required the borrower to be a current account customer. Criteria will be as much a factor for many borrowers as the deal itself.

It does highlight that there’s an arrangement fee to factor in, but borrowers need to understand there are options that can reduce or remove those costs, which could work much better for someone with a smaller mortgage.

It feels a bit mixed about what may or may not happen with rates – the bottom line is that none of us know and, as far as I know, that includes AI.

Because it was asked about two- and five-year rates, it fails to consider whether another option could fit better. An adviser would take a borrower through that, which could open up something like a three-year deal, fixing for 10-plus years and the existence and pros and cons of variable rate deals, too.

The response may give some food for thought to borrowers and help them consider the questions they want to cover with a human adviser. Most borrowers will have had a browse online of “best buy” deals and coverage of what’s happening in the market before seeking individual advice. ChatGPT may offer another way to drag a very general summary together quickly as part of that research.

Funding university

We asked: I have two children (12 and 15), both of whom are in secondary school. Both want to go to university when they finish school. What is the best way for me to start saving and how much will I need to put together? They will be going to university in the UK. Can they get student loans? How much will accommodation cost us?

ChatGPT started with details of the government loans for tuition and living costs and when they need to be repaid. It says these children would qualify for the minimum maintenance loan of between £3,907 and £6,853 depending on where they work and study.

It then said: It’s wise to start saving now with dedicated education savings. Junior Isas allow up to £9,000 a year for each child in tax-free savings or investments. Any growth is tax-free, and funds are released when the child turns 18.

Even small regular deposits (eg £50-£100/month) into a Jisa, or a high-interest children’s savings account, can add up. Other options include premium bonds, a stocks and shares Isa (higher risk, higher potential return), or simply a regular savings account in the parents’ name. Consider investments for long-term growth: index funds or diversified portfolios can beat savings rates, though with risk of loss.

It added that you should encourage children to save and to check for bursaries and scholarships.

Read ChatGPT’s full answer here

View image in fullscreen The AI chatbot did a ‘decent job of spelling out’ how much going to university could cost, says one expert. Photograph: aberCPC/Alamy

Charlene Young, a senior pensions and savings expert at AJ Bell, says: It did a decent job of spelling out the facts of how much uni could cost, and the probable level of loan that could be applied for.

It falls short on explaining the differences that would help someone make the right decision for their circumstances. A discussion with a real person would involve questions asking for extra information to give a more comprehensive answer.

When saving or investing for children, we often talk about putting on your own oxygen mask first, such as checking you have a cash buffer or emergency fund for those unexpected costs. I’d want to double check that you don’t have any unsecured, high-interest borrowing outstanding – the AI answer doesn’t really explain that, starting from an assumption your own finances are in good shape.

A pot such as a junior Isa makes sense for a specific goal such as university costs, but that money is locked away until the child turns 18 – and after that, it switches to an account in their own name that they can spend as they wish.

It was good to see the AI suggest involving the child in all financial discussions to help them become a bit self-sufficient and mitigate some of those risks, but the answer did not spell out the important access at age 18 point.

We know that investing gives the best chance of growing money over the long term – above cash and rises in the cost of living. But the short investment window between now and the time the eldest child would go to university was not mentioned at all as a risk.

A specific fund recommendation from AI would be a bit of a red flag, but it’s helpful to give people an idea of what could be considered or where to find more information. For example, what is an index fund, or a diversified portfolio and what are the likely costs.

Here, a real adviser could make a formal recommendation, or an expert could point out how to research different types of funds.

Insurance

We asked: I have just bought my first home and am expecting a child with my partner. What types of insurance do we need and how much should we spend? We travel on holidays every year and both work full time.

ChatGPT gave a list of insurances and explained what each does, including what policies typically provide and exclude. For example, it said “life insurance pays a tax-free lump sum or regular payout if a policyholder dies during the term… A Which? comparison found a 30-year-old could insure £300,000 (25-year term) for about £113 per year.”

It said the key policies to get in this case were life insurance, income protection, critical illness cover, buildings insurance, contents insurance, and travel insurance.

It said: “To save on premiums, you can bundle cover as insurers often discount multi-policy customers. A joint life policy (pays on first death) is cheaper than two single life policies, but note it ends on first claim. Sometimes buying separate dual life (first-death) and second-death life insurance can be optimal. Rates vary widely so get multiple quotes.”

Read ChatGPT’s full answer here

View image in fullscreen ChatGPT’s response on travel insurance was described as ‘very simplistic’. Photograph: Sina Ettmer/Alamy

Graeme Trudgill, chief executive of the British Insurance Brokers’ Association says the response is “very generic” and does not take into account the complexities of insurance. Brokers, he says, are able to negotiate with providers on behalf of the consumer and know what the best policies are on the market.

Kathryn Knowles from brokers Cura Insurance says some of the advice on critical illness cover was confused and did not properly distinguish it from income protection, which advisors have to do.

And Garry Nelson from AllClear Travel Insurance says the AI’s response on travel insurance is “very simplistic” and fails to help people in what they need, be it a single trip or annual policy. “This is as far as ChatGPT can go in my opinion for travel insurance enquiries – ie offering a general description as it is not regulated to offer anything more detailed or complex,” he says.

The verdict

The responses from ChatGPT were largely limited to general summaries about the various topics – and sometimes even contained inaccuracies. They may be a useful introduction and guide, but lack the level of detail needed to make serious decisions that could cost thousands of pounds. It’s a good place to start if you want an idea of some of the things you need to think about before making decisions – but if you want advice, you should be better off with a human.

* The responses from ChatGPT have been edited for this article

What to ask and how

You will get better detail in an answer if you provide a lot of information, such as your age, your salary and what debts you have, rather than just asking a general question.

Check your privacy settings to ensure you are comfortable with how ChatGPT processes your data. You may well want to ensure your chats are not being used for training.

Don’t rely on the AI to tell you about the most up-to-date products. Instead do your own research or go to an adviser. The AI is better placed to give plans or broad strategies.

Use it as a jumping-off tool, not for a final plan. While you might get some good general details, don’t make any decisions without doing your own research as well or consulting an expert.

If you don’t understand a term, ask the AI to break it down for you – for example, you might want to know what the difference is between types of loan available when you go to university.

We asked a chatbot some common finance questions – and then ran its responses past human experts

Hey AI! Can ChatGPT help you to manage your money?

Workers in the UK should turn their trepidation over AI into “exhilaration” by giving it a try or they risk being left behind by those who have, the technology secretary has said.

Peter Kyle called on employees and businesses to “act now” on getting to grips with the tech, with the generational gap in usage needing only two and a half hours of training to bridge.

Breakthroughs such as the emergence of ChatGPT have sparked an investment boom in the technology, but also led to forecasts that a host of jobs in sectors ranging from law to financial services will be affected.

However, Kyle said: “I think most people are approaching this with trepidation. Once they start [using AI], it turns to exhilaration, because it is a lot more straightforward than people realise, and it is far more rewarding than people expect.”

Kyle spoke after meeting tech company bosses to discuss a new government-industry drive to train 7.5 million UK workers – a fifth of the overall workforce – in AI by 2030, with the help of firms such as Google, Amazon and BT.

He said: “There’s no one in employment at the moment that is incapable of gaining the skills that will be needed in the economy in the next five years.

“That is the optimistic way of saying, act now, and you will thrive into the future. Don’t, and I think that some people will be left behind. And that’s what worries me the most.”

Kyle said there appeared to be a generational gap in AI, with over-55s using AI half as much as over-35s. Closing this gap would take two and a half hours of training, he said.

“People don’t need to get trained in quantum physics,” Kyle said. “They need to get trained in the basics of how AI works, how to interact with it, and to explore all of the potential it has for you as an individual in the workplace.”

Keir Starmer acknowledged this week that people were “sceptical” about AI and worry about it taking their job. Speaking at London Tech Week, the prime minister said the government would attempt to prove that technology can “create wealth in your community … create good jobs [and] vastly improve our public services”.

People in English-speaking countries including the UK, US, Australia and Canada are more nervous about the rise of AI than those in the largest EU economies, according to polling data shared with the Guardian last week.

Forecasts about the impact of AI on jobs vary, with the Organisation for Economic Co-operation and Development warning the technology could trigger job losses in skilled professions such as law, medicine and finance. The International Monetary Fund has calculated 60% of jobs in advanced economies such as the US and UK are exposed to AI and half of these jobs may be negatively affected.

However, the Tony Blair Institute, which has called for widespread adoption of AI in the public and private sectors, has said potential UK jobs losses in the private sphere will be mitigated by the technology creating new roles.

Kyle said he was ready for a reset in the debate over AI and copyright after opposition to the government’s proposed overhaul of copyright law in the House of Lords ended. The data bill, a vehicle for peers’ opposition to proposals to let AI firms use copyright-protected work without permission in order to develop their products, finally passed this week after lords did not submit further copyright-related amendments.

“I’m acting with humility and self-reflection about the things I could have done better in that process,” he said. “And I’ve made promises to move forward with a reset and a refocus on what will deliver the rights remuneration and opportunities for creatives in the digital age that they have enjoyed for generations in the analogue age – whilst travelling on that journey with the AI industry alongside.”

Peter Kyle calls on employees and businesses to act now to get to grips with technology amid forecasts of job losses

Workers in UK need to embrace AI or risk being left behind, minister says

When it comes to successful relationships, there’s nothing quite like the long, long marriage between bedbugs and humans, even if the affection goes in one direction.

The species of bedbug that feeds on us while we slumber is monogamous with humans; it does not shack up with any other species. Despite the ick factor, the insect does not transmit disease, nor does it cause harm beyond the mild irritation where its needlelike mouth pierces the skin.

That relationship, it turns out, has been going on for much longer than previously known.

According to a new study published in the journal Biology Letters, the bedbug’s long affair with humans began about 245,000 years ago. The insect strayed from the cave-dwelling bats that had been its sole source of sustenance and discovered the blood of a Neanderthal, or some other early human, that had bedded down in the same cave.

From that point on, scientists say, bedbugs diverged into two distinct species: one that lived off bat blood, and one that fed on humans.

A new study suggests that bedbugs were the first urban pest, and their population thrived in that environment. For the bloodsucking insects, it’s been the perfect 13,000-year-long marriage.

Early Humans Settled in Cities. Bedbugs Followed Them.

The starship on West 81st Street and Central Park West is losing its captain.

For nearly three decades, Carter Emmart, 64, has been director of astro-visualization at the American Museum of Natural History’s Hayden Planetarium, curating the ultimate diorama: a digital universe of data and images culled from space probes and telescopes the world over. Mr. Emmart and his crew have created a series of mesmerizing planetarium shows over the years that take audiences forward and backward in time and space to understand the origin and potential fate of the cosmos.

With shoulder-length hair, beads, bracelets and a propensity to show up for big events in a blue astronaut jumpsuit, Mr. Emmart himself seems to have been beamed in from somewhere Out There. His Manhattan office is festooned with a collection of Barbies, dolls he has used in design models, which he has lugged to all corners of the world, having dressed them in regional costumes.

To date, millions of planetarium visitors have watched and heard stars explode; galaxies collide; clouds of interstellar space dust glow, swirl and melt. They have zoomed over alien landscapes and pierced the mysterious dark matter that permeates space. The shows have been narrated by celebrities such as Tom Hanks, Robert Redford, Whoopi Goldberg and the museum’s own impresario of the sky, Neil deGrasse Tyson. The shows have been distributed to 60 different institutions in 40 countries.

A very lucky viewer could lie on the floor of the planetarium on a recent slow afternoon as Mr. Emmart led a personal tour of his digital universe, pausing to appreciate craters on the moon and the dunes of Mars.

For nearly three decades he has created mesmerizing planetarium shows at the American Museum of Natural History. But other galaxies await.

Traveling the Cosmos With Carter Emmart, One Last Time

You could be forgiven for assuming that scientists know how many kinds of proteins exist. After all, researchers have been studying proteins for more than two centuries. They have powerful tools in their labs to search for the molecules. They can scan entire genomes, spotting the genes that encode proteins. They can use artificial intelligence to help decipher the complex shapes that allow proteins to do their jobs, whether that job entails catching odors in our noses or delivering oxygen in our blood.

But the world of proteins remains remarkably mysterious. It turns out that a vast number of them have been hiding in plain sight. In a study published on Thursday, scientists revealed 4,208 previously unknown proteins that are made by viruses such as influenza and H.I.V. Researchers elsewhere have been uncovering thousands of other new proteins in bacteria, plants, animals and even humans.

Many of these newly discovered proteins probably play a vital role in life, according to Thomas Martínez, a biochemist at the University of California, Irvine. “There is no way to get around this,” he said. “If we ever want to understand fully how our biology works, we have to have a complete accounting of all the parts.”

For a long time, scientists depended on luck to find new proteins. In 1840, for example, Friedrich Ludwig Hünefeld, a German chemist, became curious about earthworm blood. He collected blood from a worm and put it on a glass slide. When he looked through a microscope, Hünefeld noticed platelike crystals: He had discovered hemoglobin.

From viruses to humans, life makes microproteins that have evaded discovery until now.

Shining a Light on the World of Microproteins

Your thumbprint, the pattern of lines in the iris of your eye: These are known to be more or less unique to each person, including you, or at least specific enough to be useful for unlocking your phone. But in a paper published Thursday in the journal Current Biology, researchers report that your patterns of breathing through your nose are so distinctive that it may be possible to identify you by breath alone, suggesting we have “breath prints.”

The study was conducted in 100 people who wore sensors for 24 hours, and the technique proved effective in distinguishing among individuals more than 90 percent of the time. The researchers who led the study also found that certain quirks of breath were linked to people’s scores on questionnaires about anxiety, among other traits, suggesting that breath monitoring over many hours may provide a useful window into mental states and disorders.

Most people rarely think about breathing, but for researchers who study smell, like Noam Sobel and his colleagues at the Weizmann Institute of Science in Israel, that regular cycle of in and out contains tantalizing information about the brain. Each inhalation comes with a firing of sensory neurons and other cells involved in monitoring the environment, and Dr. Sobel and Timna Soroka, a graduate student at the institute, wondered whether it would be possible to identify individuals from long-term recordings of their breathing patterns.

“We hypothesized, brains are unique, ergo breathing patterns would also be unique,” Dr. Sobel said.

Ms. Soroka developed a wearable sensor that fit on volunteers’ upper backs, with tubes running around to capture the airflow out of each nostril. The researchers found that by using software to analyze a day’s worth of sensor information, they could tell people apart.

Every breath you take, they really may be watching you.

Each Person Has a Unique ‘Breath Print,’ Scientists Find

For the first time, the National Hurricane Center in Miami is working with an artificial intelligence company to improve its forecasts of the powerful storms that kill thousands of people globally every year. The Atlantic season has just begun and runs through November.

DeepMind, a Google company based in London, announced on Thursday that it was supplying the government forecasters with a newly enhanced variety of its weather forecasting models. Specialized to focus on hurricanes, the model tracks a storm’s development for up to 15 days, predicting not only its path but also its strength, an ability that earlier A.I. models lacked.

Strength readings can make storm warnings far more accurate. So can reliable predictions of hurricane paths, which are known to zigzag, loop around, slow down, make hairpin turns or come to a complete stop.

The hurricane center is not eliminating its human forecasters. Instead, the Google A.I. program will be used on an experimental basis by those same experts in their existing work. Still, the research partnership is the first time in which the Miami center is drawing on an A.I. company to learn how to better warn of nature’s most destructive storms.

The National Hurricane Center will experiment with the company’s DeepMind program to enhance the work of its expert meteorologists.

Google DeepMind and National Hurricane Center Join on New A.I. Model

First Close-Up of Star Outside Our Galaxy Shows a Giant About to Blow

Astronomers zoomed in on a stellar behemoth in the Larger Magellanic Cloud, a galaxy that orbits about 160,000 light-years from the Milky Way.

By

Breaking news, videos and photos on space and astronomy, including NASA, SpaceX, black holes, the moon, Mars, Jupiter, Saturn, the International Space Station, the sun, exoplanets and the stars.

Space and Astronomy

Australia must “lean in hard” to the benefits of artificial intelligence or else risk ending up “on the end of somebody else’s supply chain”, according to the new industry and science minister, Tim Ayres, with the Labor government planning to further regulate the rapidly evolving technology.

Ayres, a former official with the manufacturing union, acknowledged Australians remained sceptical about AI and stressed that employers and employees needed to have discussions about how automation could affect workplaces.

The minister said Australia had “no alternative” but to embrace the new technology and seek to become a world leader in regulating and using AI.

“It’s the government’s job to lean into the opportunity to outline that for businesses and for workers, but also to make sure that they are confident that we’ve got the capability to deal with the potential pitfalls,” Ayres told Guardian Australia.

“I think the Australian answer has got to be leaning in hard and focusing on strategy and regulation that is in the interest of Australians.”

Ayres was promoted to cabinet last month after holding the junior portfolio of assistant minister for manufacturing and trade last term.

Now as minister for industry, innovation and science, Ayres has greater direction of the Labor government’s flagship Future Made In Australia initiative – the umbrella plan tying together policies around manufacturing, the energy transition, research and business.

One of Ayres’ more public-facing and immediate challenges, however, is around AI policy. His predecessor, Ed Husic, set up several processes around the development of local industry and mandatory guardrails around its use, including discussion of a new standalone AI act and legislation.

Ayres, who has been minister for less than a month, says the government is still setting its course of action, considering approaches from like-minded countries and the breakneck speed of the technology’s evolution. He said the response would include legislation and regulation, which was still to be decided, but that Australia would benefit from moving quickly.

“There is no alternative to the approach of Australia leaning in,” he said. “That’s how we’re going to shape a digital future for Australia. That’s how we get some agency and say over the way that the technology develops and our global partners develop on these questions.

“The alternative is to sit back right and just to be on the end of somebody else’s supply chain.”

The minister said there would be “a lot to gain” from embracing AI, particularly highlighting potential boosts to productivity and economic growth. Ayres, who grew up on a cattle farm near Lismore, said both white-collar and traditional blue-collar jobs could reap benefits from adopting automation and new technology.

He said that his experience in the manufacturing union was that many workers accepted “the cold reality that the only thing more disruptive in the labour market is standing back and allowing Australia to become a cul-de-sac in technological terms”.

“But I do want to see firms and employers being thoughtful about making sure that we’re adopting AI in a way that makes jobs better,” Ayres said.

“We’ve got an industrial relations framework that provides for proper consultation and engagement at the firm level … there’s going to be discussions workplace by workplace, around these questions.”

Last week, the Business Council of Australia released a major report calling for Australia to become a global leader in AI, talking up its potential to boost productivity and lift living standards through economic growth.

The Australian Council of Trade Unions in December warned one in three Australian workers are at risk of job loss from the introduction of AI.

“The BCA paper released last week pointed to a high level of scepticism amongst Australians about this new wave of technology. That’s not unusual for Australia,” Ayres said.

“Every wave of technological change reshapes the labour market. That’s the truth of it. The negative consequences in jobs terms of technological change have historically been outweighed by the new investment and the new developments in jobs and technologies.”

Ayres also said Labor would continue with its Future Made In Australia agenda, which would include an “aggressive” focus on critical minerals, iron and steel production and boosting manufacturing as part of the renewable energy transition.

“I want to make sure that we’re doing everything we can to be building new factories and industrial capability,” he said.

“Particularly in regions like central Queensland and the Hunter and La Trobe, where the intersection of our future energy advantage and our industrial capability really means Australia can deliver, and deliver for those communities as well.”

Tim Ayres says the Albanee government will focus on legislation and regulation but that Australia would benefit from moving quickly

Australia has ‘no alternative’ but to embrace AI and seek to be a world leader in the field, industry and science minister says

No Antarctic gonate squid had ever been seen alive before, as far as the team was aware. They followed it for a couple of minutes and made sure to record it on video, capturing the creature’s red coloration and white spots.

“Videos like this get me really excited,” said Linsey Sala, a museum scientist who manages the pelagic invertebrate collection at Scripps Institution of Oceanography and was not involved in the expedition. Discoveries of species like this “can be really informative to how they live life at great depths,” Ms. Sala said. Unidentified specimens might be sitting in collections around the world, she added, in which case the video footage could be helpful in revealing what they are.

Previous sightings of this species of squid have been limited to individuals caught by fishing vessels and squid remains found in other marine animals, mostly in the Falkland Islands.

“It’s always exciting to see live footage of a critter that was known only from dead specimens previously,” said Bruce Robison, a deep sea ecologist at the Monterey Bay Aquarium Research Institute who also was not involved in the expedition.

For those aboard the vessel, the discovery was “sheer elation,” Dr. Thurber said. There was a fever of excitement even before they had properly identified the squid. To verify the species, Dr. Thurber and his colleagues sent the images captured by the underwater vehicle to taxonomists around the world.

An expedition in the Southern Ocean captured video of a rare species of deep-sea cephalopod. Until now, it had been found only in fishing nets and in the bellies of seabirds.

This Elusive Antarctic Squid Was Seen for the First Time

Spacecraft and specialized telescopes have been studying the sun closely for decades, probing the secrets of its spots, flares and corona. But neither human nor robotic eyes had properly seen the north or south poles of the sun.

On Wednesday, the European Space Agency released the first clear images of the sun’s south pole, which were captured in late March by its Solar Orbiter spacecraft.

“It’s the first time ever that humanity has had an image of the poles of the sun,” said Carole Mundell, director of science for the European Space Agency. “It’s a wonderful achievement.”

Scientists have had vague glimpses of the sun’s poles before. However, those side-on views have been akin to “looking through grass,” said Mathew Owens, a space physicist at the University of Reading in England, who added that it has been difficult to work out what is happening at the poles and how they differ from other regions of the star.

Visuals from the European Space Agency’s Solar Orbiter reveal chaotic solar magnetism in the solar polar region. Even better images are expected in the years ahead.

First Ever Images of Sun’s South Pole Released by Solar Orbiter

Elon Musk has expressed contrition for some of his tweets about Donald Trump last week, in an apparent effort to retreat from an explosive falling out that has threatened to damage the Tesla boss’s business interests.

Musk was by far the biggest donor to Trump’s presidential campaign, but tensions between the two erupted into public view last week and rapidly escalated, as the world’s richest man called for the president’s impeachment and mocked his connections to the sex offender Jeffrey Epstein in a series of posts.

On Wednesday, Musk posted on X, the social network he owns: “I regret some of my posts about President @realDonaldTrump last week. They went too far.”

Asked about the apparent apology, the president told the New York Post: “I thought it was very nice that he did that.” The comments came after the paper published a previously recorded interview, in which Trump had said “I guess I could” reconcile with Musk.

The president said he “was not a happy camper” when the Tesla chief executive launched last week’s tirade but claimed he had “no hard feelings for it”, adding: “I think he feels very badly, that he said that.”

The possibility of a rapprochement – however superficial – appeared to be welcomed by investors. Tesla’s share price rose by 2.6% in pre-market trading.

The public feud was one of the most extraordinary turns in the relationship between the two men. During the presidential campaign they had claimed to be ideological allies, and Musk briefly served in Trump’s government as head of the so-called “department of government efficiency”, a drive to slash government programmes nicknamed “Doge” after the internet meme. Experts have argued that the cost-cutting is illegal.

However, relations quickly soured after Musk publicly criticised Trump’s “big beautiful bill”, saying it would add $2.4tn to US government borrowing and calling it a “disgusting abomination”.

Trump had responded to Musk’s barrage of criticism by saying the tech billionaire “went crazy”. However, the president also made direct references to Musk’s companies, highlighting the potential financial risks of the feud.

In a direct reference to Tesla, Trump wrote on Truth Social, his own social media network: “I took away his EV Mandate that forced everyone to buy Electric Cars that nobody else wanted.” The electric car pioneer has been struggling with falling sales in some markets, including much of Europe, in part, analysts say, because of Musk’s allegiance with Trump.

Musk’s alignment with the US president had prompted a surge in the market value of Tesla, with investors hoping that the White House would look more favourably at the company’s autonomous driving technology. Musk’s retreat from the feud came a day before Tesla’s launch of a “robotaxi” service in Austin, Texas. That launch is seen as crucial for the company to justify its position as the world’s most valuable carmaker even as it struggles with an ageing lineup of products.

Trump also threatened another of Musk’s key businesses, the rocket company SpaceX. Trump wrote: “The easiest way to save money in our Budget, Billions and Billions of Dollars, is to terminate Elon’s Governmental Subsidies and Contracts.”

In practice the US government is unlikely to cancel SpaceX’s contracts, as it carries out more strategically important satellite launches than every other company in the world combined. Musk at first threatened to decommission SpaceX’s Dragon spacecraft, the key vehicle to transport Nasa astronauts to the International Space Station, before withdrawing the threat.

Tesla share price rises as former Doge head seems to retreat from feud and president hints at reconciliation

‘They went too far’: Musk says he regrets some of his posts about Trump

Before Ibtihal Aboussad was fired by Microsoft for protesting the company’s work with the Israeli military during a celebration of the firm’s 50th anniversary, she sent two emails.

The first went to all of her colleagues. She appealed to their universal humanity and urged them to stand against Microsoft’s contracts to provide cloud computing software and artificial intelligence products to the Israeli Defense Forces (IDF).

She sent the second to the “Muslims at Microsoft” email list. Its subject line read: “Muslims of Microsoft, Our Code Kills Palestinians.”

With her email, Aboussad told the Guardian, she wanted Muslim staff of companies such as Microsoft, Google and Amazon to stop regarding the question of whether they organize against their employer’s work with the IDF as an issue of secular or professional ethics. It was a question of Islam, of their faith, she argued.

“I wanted to say, ‘Hey, remember, your rizq [livelihood] is from Allah,’” Aboussad said. “It should be clean, and you cannot be contributing to oppression.”

There’s been protests within American tech companies against their contracts with Israel and its military since the start of the war in Gaza. There have been walkouts and vigils. Offices have been taken over and op-eds have been written. Some staff have resigned and some have been fired for their activism.

But as the war endures and Palestinians in Gaza are being starved, forcibly displaced and killed, and the contracts have survived, there’s a growing group of Muslim staffers who are unsure whether they can religiously justify working at companies that they view as effectively defense contractors.

View image in fullscreen Ibtihal Aboussad participates in a protest in Seattle on 4 April 2025. Photograph: No Azure for Apartheid

The Guardian spoke with nearly a dozen Muslim employees of major tech companies who’ve been grappling with the question, many of whom asked not to be named for fear of professional repercussions or because they are continuing to organize inside the companies.

Several of the staffers have resigned or are in the process of resigning. Some said they aren’t convinced that their employers have crossed a line that would force them to quit. Others worried that leaving the company would do more harm than good.

Nearly all said that the public protest of Aboussad, and her co-demonstrator Vaniya Agrawal, has pressed the question. And at least for some, the footage of Aboussad – a hijab-clad young woman staring down Microsoft’s AI CEO, Mustafa Suleyman, and calling him a war profiteer – has been activating.

Organizers at No Azure for Apartheid (NoAA), the worker-led group at Microsoft that Aboussad belongs to, and whose name refers to Microsoft’s cloud-computing platform, say that they’ve heard from more staffers looking to get involved with the group since Aboussad’s action. And they’ve heard from more than a dozen workers looking to resign or lodge a protest, and several more who have already done so.

Others are on the fence, with some in a constant spiritual crisis over whether they should resign. The reasons vary: some feel they can do more to agitate against these contracts from inside the company, others worry about being able to provide for their families and some worry about trying to find another tech job in a tough market.

“Honestly, I’ve been praying about what Allah wants me to do,” one Microsoft employee said. “Because it doesn’t seem like it’s right for a Muslim to continue working for such companies. But if we leave, then there could be a pro-Israeli person who takes our spot, and then you’re not serving the cause by doing this.”

One Google worker said they decided it no longer feels halal to remain at the company. But their father doesn’t agree: he argues it’s their “Islamic duty” to stay at the company because the role was a blessing from God. “He said this is self-sabotage,” the worker said.

View image in fullscreen Members of No Azure for Apartheid hold a protest on 20 March 2025. Photograph: No Azure for Apartheid

“My parents are like, ‘If you quit, how does that help the cause?’ I don’t think they understand that by quitting I am not trying to help the cause,” the Google staffer said. “I’m at a [spiritual and moral] negative right now. I’m just trying to go back to neutral. Every day my work is actively harming people. I can’t help people if I’m actively harming them.”

The debate is not confined to the cafeterias of tech companies. Several tech workers said they have sought the spiritual insight of popular Islamic scholars, including Imam Omar Suleiman, the founding president of the Yaqeen Institute, a Texas-based Islamic research organization. Suleiman is actively engaging with Microsoft workers to help determine “if there’s any part of Microsoft that wouldn’t be considered complicit in this”, he said. The imam said he is still grappling with the question of whether and how urgently Muslims need to quit their tech jobs, leaving the workers to wrestle with it privately for now.

“More fields of employment are complicated than not,” Suleiman said. “It’s not always as straightforward as someone that works at a liquor store. It’s someone that works at a grocery store [that sells alcohol].”

Big tech’s rising role in the siege of Gaza

Details of how big tech works with the IDF have long been murky, and many tech staffers had mostly accepted their employers’ denials or defenses of these contracts. But recently, evidence that the tech industry’s products have been used in Israel’s violent campaign in Gaza, which the UN has concluded is consistent with “the characteristics of genocide”, has been mounting.

Microsoft deepened its ties with the Israeli military in the wake of Hamas’s 7 October 2023 attacks, according to a Guardian investigation reported in collaboration with +972 Magazine and Local Call, based in part on documents obtained by Drop Site News. Leaked documents indicated Microsoft has a “footprint in all major military infrastructures” in Israel. The Associated Press has also reported that Microsoft technology has aided in Israel’s surveillance of Palestinians.

Microsoft has defended its contract with the IDF, saying that an internal investigation concluded the firm “found no evidence” that its technology was used by the IDF to target or harm people.

Microsoft’s chief communications officer, Frank Shaw, reiterated those findings and said these reports “are not accurate”.

“As we stated before in our blog, we have found no evidence to date that Microsoft’s Azure and AI technologies have been used to target or harm people in the conflict in Gaza,” he said in a statement.

View image in fullscreen Hossam Nasr participates in a protest in 2024. Photograph: Courtesy of Hossam Nasr

The company’s investigation did nothing to assuage workers. Hossam Nasr, an organizer with NoAA who was fired by Microsoft after organizing a vigil for Palestinians in 2024, argued there is no way to have an “ethical” contract with a military “whose leaders are wanted by the international criminal court for war crimes”.

Google and Amazon both provide cloud services to the Israeli military and government under a $1.2bn agreement dubbed “Project Nimbus”.

Google has maintained that its technology is not aimed at military work, but reporting suggests the tech giant provides cloud services as well as advanced AI and machine learning tools that directly equip the Israeli military with various features including image and object detection and analysis. Google and Amazon did not respond to requests for comment.

‘If you do not organize, you must leave’

Many of the tech workers the Guardian spoke with said the recent revelations about the depth of their employers’ work with the IDF and the lack of sufficient response to worker opposition to the contracts felt like a mask-off moment, one that left little doubt about their employers’ roles in Israel’s offensive and even less hope that the companies might stop working with the Israeli military.

In April, the Boycott, Divest and Sanction (BDS) movement, which seeks to use economic means to protest Israel’s occupation and siege of the Palestinian territories, added Microsoft to its list of companies to boycott, further pressing the issue for many Muslim tech workers.

It was ultimately these revelations that moved Aboussad to stage her protest at Microsoft’s 50th anniversary event in April. But there were other factors: in addition to becoming involved with NoAA in February 2025, Aboussad had also started wearing hijab at the end of 2023, and she felt a new responsibility as someone who was more visibly Muslim to represent her faith appropriately and stand in opposition to the company’s contract with the IDF.

In her email to Muslim staff, she included an essay written by Hasan Ibraheem, one of several Google workers fired for occupying the company’s New York office the year before. The essay was meant to serve as a “dire callout to our community”, Aboussad wrote in the email.

“To my Muslim brothers and sisters, I offer this essay as a sincere naseeha [advice],” Ibraheem’s essay began. Then he quoted from the prophet Muhammad (peace be upon him):

“Whosoever of you sees an evil, let him change it with his hand; and if he is not able to do so, then with his tongue; and if he is not able to do so, then with his heart – and that is the weakest of faith.”

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Ibraheem’s message was simple: Muslims have a religious obligation to stop oppression wherever they see it. Working at a company that contracts with the Israeli government as it continues its decimation of Gaza without doing anything to push back against those contracts was in violation of that obligation. Put bluntly, Muslims have two options, Ibraheem said: to either fight or quit.

“If you do not organize, you must leave,” Ibraheem wrote. “And even if you organize, your goal should be to eventually leave. Organizing does not absolve you of complicity indefinitely.”

View image in fullscreen Hasan Ibraheem. Photograph: Ahmed Gaber/The Guardian

“If you know that the company you work for is directly enabling harmful activities, then maybe your money is not completely halal,” Ibraheem said.

Aboussad said that in the days after she protested in the meeting, she received dozens of direct messages on Instagram from other Muslim tech workers. Many said she inspired them to do more to oppose their company’s actions, and some said they were even thinking of resigning, Aboussad recalled. One person said her demonstration “removed any excuses” they might have made about not doing more to oppose their employers’ work with the IDF.

One confrontation inspires multiple resignations

The debates aren’t limited to the companies’ US offices. For some workers at Microsoft’s offices across the Middle East and north Africa, Aboussad and Agrawal’s protest was one of the first times they were confronted with the depth of their employers’ work with the Israeli military. For others, their concerns about their roles at the company had long been brewing. One worker, based out of Microsoft’s Cairo office, told the Guardian she had been weary for months of what she saw as a vehement pro-Israeli stance in Microsoft’s internal communications to employees and their lack of mention of the now more than 50,000 Palestinians who have been killed by the IDF.

The first time she began to wonder whether she belonged at Microsoft as a Muslim was when the company fired Nasr and another worker after the two organized a vigil for Palestinians.

Aboussad and Agrawal’s protest about six months later helped her answer that question, she said. Their demonstration prompted about 100 employees in the Cairo office to take a day off in protest of Microsoft’s work with Israel – an action that was just shy of striking, which is generally illegal in Egypt. On that day, she decided she would quit.

“The response of [Suleyman] and [Microsoft CEO] Satya [Nadella] was very dismissive,” she said. “And I think Satya laughed, and that sort of made me feel like, no, I don’t think I belong here, and me staying here is just supporting what they’re doing.”

None of the Google or Microsoft workers who spoke to the Guardian had any doubt about whether their employers were contributing to Israeli military operations. But some held out hope the company may switch strategies.

One person who has worked at Microsoft for nearly a dozen years said they felt betrayed by the company, which sold them on its 2014 mission of “empowering every person and every organization on the planet to achieve more”. But they remain hopeful that internal pressure might force the company to change its policy on working with the Israeli government.

“I know during South African apartheid, the company did flip to the right side,” they said, citing Microsoft’s decision to leave South Africa in 1986 in response to the country’s laws enforcing racial discrimination. “I hope that history repeats itself and they become the company where the culture is that you make everyone feel great.”

Meanwhile, workers’ protests have continued. At Microsoft, NoAA has disrupted at least two more company events, prompting the company to fire at least one of the workers, Joe Lopez.

Another staffer, senior UX designer Jasmina Mathieu, resigned publicly, stating in an email to leadership and employees that she could no longer work at a company that directly or indirectly enables “horrific actions done by Israel”. Since then, the Microsoft staffers said, they can no longer send emails to anyone, including human resources, that contained the words “genocide”, “Gaza”, “Palestine” or even “Vaniya Agrawal”. One worker with NoAA, Nisrine Jaradat, was able to get around the block and send an email to all Microsoft staff decrying the policy that “utterly and completely discriminate[s] against an entire nation, an entire people, and an entire community”. Jaradat called on her co-workers to either organize against the company’s work with the IDF or quit.

“If you choose to leave Microsoft to no longer be com​plicit in genocide do not go quietly,” Jaradat wrote.

A popular imam fields the question: is working in big tech haram?

Many workers have consulted with local religious leaders and scholars but have not been given a clear response. “We’ve been wanting to get an opinion that says, you know, ‘Just quit, you should not be working here,’ but we did not,” one Microsoft worker said. “They just say it depends on the situation.”

Imam Omar Suleiman has been helping people figure out whether working at certain companies is halal or permissible for the entire 20 years he’s been a religious leader.

Over the last two years, Muslims tech workers from around the country have reached out to Suleiman’s Yaqeen Institute with the same question: can I still work at a tech company that is helping power the Israeli military? For Suleiman, tech has been one of the hardest fields to navigate “because you have Muslims that work at various levels and tech companies are involved in this genocide to varying extents”, he said.

View image in fullscreen A protester waves a Palestinian flag outside the Seattle Convention Center on 4 April 2025. Photograph: Andres Robles/No Azure for Apartheid

What Suleiman and the Yaqeen Institute – where Aboussad now works – ultimately decide matters a great deal to tech workers of a certain age. The institute is well regarded among millennial Muslims in the US for its easy-to-digest content on how to understand and apply the teachings of the Qur’an. Suleiman’s videos have garnered him more than 3 million Instagram followers. Nearly every tech worker who the Guardian spoke to cited Suleiman’s speeches shared on social media as inspiration for why they want to leave Google or Microsoft. Yet for Suleiman, who recently gave an impassioned speech, or khutbah, at a Virginia mosque urging congregants to “have some dignity” and quit their jobs at arms manufacturers, whether Islam requires that these tech workers leave their jobs with the same urgency is still an open question. Like Suleiman’s example of a clerk at a grocery store selling alcohol alongside fruits and vegetables, it is not as clearcut if coding productivity tools poses the same grave transgression as building a bomb.

Suleiman and others at the Yaqeen Institute are working on developing a general framework that will help Muslims decide whether they can religiously justify working at any of the tech firms contracting with the IDF. That framework will be published as a resource for both local imams as well as individuals in the midst of a spiritual crisis about their jobs.

Suleiman looks at cases from many different sides. Staff have to ask themselves personal questions. Islamic religious law advises all Muslims to leave something that causes them doubt for something that doesn’t. But, he cautions, many cases are complicated. You might be the primary breadwinner of your household, have an employer-sponsored visa or be working at a part of the company that you feel doesn’t have anything to do with the products the firm builds for militaries, such as LinkedIn. One also has to evaluate whether you’re having a positive impact by staying in the company.

“There’s room for the person who holds back the hand of the pharaoh from inside the pharaoh’s court,” Suleiman said. “But they have to demonstrate how exactly [they are] minimizing that harm without at any point becoming a mouthpiece for oppression.”

And they need to evaluate the workings of their companies, he said. Do they produce haram, or impermissible, products, like alcohol or weapons used to kill people? In those cases, “they need to leave their job and they need to find a different job depending on what layer they participate in and how much need they have”, Suleiman said. Or is the company based around usury or interest-based transactions, which are also impermissible in Islam? In those cases, you can still work at the company, but you should up your charity to offset those haram transactions.

While Islamic jurisprudence has an established structure to make these decisions, Suleiman says tech remains a complicated matter to locate within that framework. The imam said he is still in the discovery phase on how the roles workers find themselves in at these companies align with their religious duty.

“It’s very hard to compare something that’s just purely generating weapons of mass destruction and a tech company that has a wide array of businesses, but also happens to be manufacturing for a genocide,” he said. “It’s hard for me to figure out when that line disappears.”

Is working in big tech halal? Muslim workers are reckoning with the possibility that their jobs go against their religious obligations

As big tech grows more involved in Gaza, Muslim workers are wrestling with a spiritual crisis

Meta is to announce a $15bn (£11bn) bid to achieve computerised “superintelligence”, according to multiple reports.

The Silicon Valley race to dominate artificial intelligence is speeding up despite the patchy performance of many existing AI systems.

Mark Zuckerberg, Meta’s chief executive, is expected to announce the company will buy a 49% stake in Scale AI, a startup led by Alexandr Wang and co-founded by Lucy Guo, in a move described by one Silicon Valley analyst as the action of “a wartime CEO”.

Superintelligence is described as a type of AI that can perform better than humans at all tasks. Currently AI cannot reach the same level as humans in all tasks, a state known as artificial general intelligence (AGI).. Recent studies have shown that many mainstream systems collapse when presented with highly complex puzzles.

Meta’s attempt to leapfrog the current state of progress and target superintelligence is seen by observers as an attempt by the company to regain the initiative over AI after significant advances by competitors including Sam Altman’s OpenAI and Google and after Meta’s huge investment in the concept of the Metaverse flopped..

In March, Wang, 28, signed a deal to build a US defence department system called ThunderForge to apply AI to US military planning and operations, starting with the Indo-Pacific and European commands. The company also received early investment from Peter Thiel’s Founders Fund.

The move by Meta has led to fresh calls for European governments to launch their own more transparent research push that would ensure that the technology was developed robustly while maintaining public trust – an AI equivalent of the Cern European Organization for Nuclear Research in Switzerland.

Michael Wooldridge, a professor of the foundations of artificial intelligence at the University of Oxford, said: “There’s a good argument that there should be a Cern for AI where governments collaborate to develop AI openly and robustly so we understand and trust the technology we are developing. That’s not going to happen if it’s developed behind closed doors. [AI] seems just as important as Cern and particle accelerators.”

Wooldridge said the reported deal “seems very much like an attempt by Meta to regain the initiative after the Metaverse didn’t take off. They invested spectacular amounts in that and it didn’t just fail, it was ridiculed.”

But he said the state of AI development was patchy and that AGI is still far away, never mind “superintelligence”.

“We have AI that can do genuinely impressive things but then it fails on an absolutely simple task which any competent GCSE student wouldn’t fail,” he said.

Dr Andrew Rogoyski, the director of partnerships and innovation at the Institute for People-Centred AI at Surrey University, said the deal would be “a reflection that the AI companies are hoovering up AI talent from wherever they can get it, whether acquiring young AI companies, or hollowing out university AI departments”.

He said that Meta’s strategy for AI was different from the likes of OpenAI and Anthropic in that AI is an enabler for Meta’s business rather than its core purpose.

“It means they’re not quite as desperate to achieve AGI, so they can afford to take a longer view,” he said.

It was reported that Wang will take a senior position in Meta.

Meta declined to comment. Scale AI has been contacted for comment.

Mark Zuckerberg expected to announce Meta will buy 49% stake in Scale AI as race to dominate AI market speeds up

Meta to announce $15bn investment in bid to achieve computerised ‘superintelligence’

Disney and Universal sued an artificial intelligence company on Wednesday, alleging copyright infringement. In their lawsuit, the entertainment giants called Midjourney’s popular AI-powered image generator a “bottomless pit of plagiarism” for its alleged reproductions of the studios’ best-known characters.

The suit, filed in federal court in Los Angeles, claims Midjourney pirated the libraries of the two Hollywood studios, making and distributing without permission “innumerable” copies of their marquee characters such as Darth Vader from Star Wars, Elsa from Frozen, and the Minions from Despicable Me. Midjourney did not immediately respond to a request for comment.

The suit by Disney and Universal over images and video represents a new frontier in the raging legal wars over the copyright and the creation of generative artificial intelligence. Previous suits have covered copyrighted text and music; Disney and Universal are two of the biggest industry players thus far to sue over images and videos.

Horacio Gutierrez, Disney’s chief legal officer, said in a statement: “We are bullish on the promise of AI technology and optimistic about how it can be used responsibly as a tool to further human creativity, but piracy is piracy, and the fact that it’s done by an AI company does not make it any less infringing.”

NBCUniversal’s executive vice-president and general counsel, Kim Harris, said the company was suing to “protect the hard work of all the artists whose work entertains and inspires us and the significant investment we make in our content”.

The studios claim the San Francisco company, one of the earliest to offer an AI product that generated images, rebuffed their request to stop infringing on their copyrighted works or, at a minimum, take technological measures to halt the creation of AI-generated images that copied the characters.

Instead, the studios argue, Midjourney continued to release new versions of its AI image service that boast higher-quality infringing images. Midjourney can recreate animated images from a typed request, or prompt. AI companies use enormous troves of data, often retrieved from millions of websites, to train programs to generate new images or text.

In a 2022 interview with Forbes, Midjourney CEO Holz said he built the company’s database by performing “a big scrape of the Internet”.

In the suit filed by seven corporate entities at the studios that own or control copyrights for the various Disney and Universal Pictures film units, the studios offered examples of Midjourney animations that include Disney characters, such as Yoda wielding a lightsaber, Bart Simpson riding a skateboard, Marvel’s Iron Man soaring above the clouds and Pixar’s Buzz Lightyear taking flight. The image generator also recreated such Universal characters as How to Train Your Dragon’s dragon, Toothless, the green ogre Shrek and Po from Kung Fu Panda.

“By helping itself to plaintiffs’ copyrighted works, and then distributing images (and soon videos) that blatantly incorporate and copy Disney’s and Universal’s famous characters – without investing a penny in their creation – Midjourney is the quintessential copyright free-rider and a bottomless pit of plagiarism,” the suit alleges.

Disney and Universal asked the court for a preliminary injunction, to prevent Midjourney from copying their works, or offering its image- or video-generation service without protections against infringement. The studios also seek unspecified damages.

Midjourney, founded in 2021 by David Holz, monetizes the service through paid subscriptions and generated $300m in revenue last year alone, the studios said.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

This is not the first time Midjourney has been accused of misusing artists’ work to train their AI systems. A year ago, a California federal judge found that 10 artists behind a copyright infringement suit against Midjourney, Stability AI and other companies had plausibly argued these AI companies had copied and stored their work on company servers, and could be liable for using it without permission. That ruling allowed the lawsuit over the unauthorized use of images to proceed. It is in the process of litigation.

The cases are part of a wave of lawsuits brought by copyright owners including authors, news outlets and record labels against tech companies over their use of copyrighted materials for AI training without permission.

Asked whether he sought consent of the artists whose work was covered by copyright, Holz responded: “there isn’t really a way to get a hundred million images and know where they’re coming from.” OpenAI said early last year in a filing to the UK government that it would be “impossible to train today’s leading AI models without using copyrighted materials”.

In late 2023, the New York Times sued OpenAI, maker of ChatGPT, and Microsoft, which owns a 49% stake in the startup, alleging unauthorized use and reproduction of its articles, which comprise a vast corpus of text. The suit is ongoing. Other news outlets, including the Guardian, have reached agreements to license their archives to artificial intelligence companies. Book authors have likewise sued Meta for alleged use of a huge database of pirated books to train its Llama AI models, though many of the writers’ claims have been dismissed.

In June 2024, major record labels sued two AI companies for copyright infringement. Sony Music Entertainment, Universal Music Group Recordings and Warner Records accused Suno and Udio of exploiting millions of songs to create engines that could “spit out” obviously derivative music.

Studios accuse AI firm of ‘piracy’ and seek injunction over alleged use of copyrighted characters

Disney and Universal sue AI image creator Midjourney, alleging copyright infringement

The use of ChatGPT and other artificial intelligence tools are becoming increasingly commonplace in UK higher education. A February survey of 1,000 students showed an “explosive increase” in use of generative AI in particular over the previous 12 months.

With this in mind, we’d like to find out more about how AI is affecting students at university.

How has AI impacted your studies? Have you used AI tools? Have you been suspected of using AI when you haven’t? What guidance have you been given by universities or tutors about using AI? Do you have any concerns?

We’d also like to hear from university teaching staff – what is the impact of AI on students’ work? What are the challenges?

Share your views and experiences

You can tell us about your experiences with AI at university using this form.

Please share your story if you are 18 or over, anonymously if you wish. For more information, please see our terms of service and privacy policy.

Your responses, which can be anonymous, are secure as the form is encrypted and only the Guardian has access to your contributions. We will only use the data you provide us for the purpose of the feature and we will delete any personal data when we no longer require it for this purpose. For true anonymity please use our SecureDrop service instead.

If you’re having trouble using the form, click here. Read terms of service here and privacy policy here.

We would like to hear from university students and teaching staff about the impact of AI on studies

UK students and staff: tell us your experiences with AI at university

Having been instrumental in the reduction of smartphones to metal and glass slabs devoid of distinguishing features, Samsung hopes that going thinner and lighter with a special Edge edition of its high-end Galaxy S25 Android will prove design innovation isn’t dead.

The Guardian’s journalism is independent. We will earn a commission if you buy something through an affiliate link. Learn more.

The S25 Edge is very thin at just 5.8mm thick – if you ignore the camera bump on the back – making it a full 1.5mm thinner than its similarly sized S25+ sibling and about the same thickness as a stack of seven credit cards. Its light 168g weight makes it feel even thinner than the numbers suggest and photos don’t do it justice.

The extra thinness costs about £100 more than Samsung’s other 6.7in phone with the S25 Edge priced from £1,099 (€1,259/$1,099/A$1,849) placing it between the £999 S25+ and £1,249 S25 Ultra.

View image in fullscreen Despite its super-slender frame, the S25 Edge looks like any other large Samsung from the screen side. Photograph: Samuel Gibbs/The Guardian

The large high-quality 6.7in Oled screen is bright, crisp and smooth with a rapid fingerprint scanner embedded in it. The phone’s light weight makes it easier to hold than other big handsets but it still requires two hands to use the majority of the time. Slap it in a case and the thickness advantage is essentially removed, so the Edge is a phone best used without one.

The rest of the phone is very similar to other Samsungs. It has Qualcomm’s top Snapdragon 8 Elite chip, plenty of Ram and at least 256GB of storage. It flies through apps, games or anything else you might want to do with a phone. Samsung’s version of Android 15 (One UI 7) works well with plenty of customisation options, though some of the default settings make it a lot more like Apple’s iOS than you’d expect.

It is packed with plenty of Google and Samsung’s most advanced AI tools. Some of them are great, such as Gemini; some are handy occasionally, such as the image editing tools, and others can safely be ignored. The Now bar and live notifications, which show ongoing tasks such as music playing in Spotify, are super handy, appearing on the lock screen and in the task bar on the home screen.

Samsung will provide the Edge with software updates until 31 May 2032, making it one of the longest supported phones available.

View image in fullscreen The camera cluster sticks out the back of the S25 Edge for almost the full thickness of the phone again. Photograph: Samuel Gibbs/The Guardian

Specifications

Screen: 6.7in QHD+ Dynamic Amoled 2X 120Hz

Processor: Qualcomm Snapdragon 8 Elite for Galaxy

RAM: 12GB

Storage: 256 or 512GB

Operating system: One UI 7 (Android 15)

Camera: 200MP + 12MP 0.6x; 12MP front-facing

Connectivity: 5G, USB-C, wifi 7, NFC, Bluetooth 5.4, UWB and GNSS

Water resistance: IP68 (1.5m for 30 mins)

Dimensions: 158.2 x 75.6 x 5.8mm

Weight: 163g

View image in fullscreen The S25 Edge takes about 80 minutes to fully charge, hitting 50% in 26 minutes using a 25W or greater USB-C power adaptor (not included). Photograph: Samuel Gibbs/The Guardian

The Edge’s extreme slender design does come with a few drawbacks, the biggest of which is a small battery with a slightly lower capacity than the regular S25. As a result, despite being a large phone the Edge has fairly short battery life. It will last about a day and a half of light usage between charges, so should see out most days, but it falls just short of the smallest S25 and about a day behind the the S25 Ultra.

It also gets a little hotter than other Samsung models when gaming, but if you play a lot of Fortnite or similar games the small battery capacity is probably already a cut too far.

Camera

View image in fullscreen The Samsung camera app is fairly easy to use, while the light weight of the S25 Edge helps with one-handed photography. Photograph: Samuel Gibbs/The Guardian

The other big trade-off for the svelte frame is in the camera department. It only has two cameras on the back, lacking a telephoto sensor as fitted to the rest of the S25 series.

The main camera is an excellent 200-megapixel model that is very similar to that on the S25 Ultra. It is a top-drawer camera that is better than the 50MP main cameras on the S25 and S25+. It shoots great photos across a range of conditions, and manages a very good 2x in-sensor zoom to somewhat negate the lack of a dedicated telephoto camera. The 12MP ultra wide camera is solid and can be used for macrophotography, while the selfie camera is very good.

The camera app has the same long list of modes as other S25 models, shoots great video and manages to be fairly easy to use. Overall the main camera is excellent and if you never wish to close the distance to objects with a real zoom, it might be all the camera you need.

Sustainability

View image in fullscreen The titanium frame and hardened glass feel extremely solid despite the phone’s svelte dimensions. Photograph: Samuel Gibbs/The Guardian

Samsung does not provide an expected lifespan for the battery but it should last in excess of 500 full charge cycles with at least 80% of its original capacity.

The phone is generally repairable. Screen repairs cost £259 by authorised service centres and include a battery replacement. Samsung also offers a self-repair programme.

The phone contains recycled materials, while Samsung offers trade-in and recycling schemes for old devices. The company publishes annual sustainability reports and impact assessments for some individual products.

Price

The Samsung Galaxy S25 Edge costs from £1,099 (€1,259/$1,099/A$1,849).

For comparison, the Galaxy S25 Ultra costs £1,249, the S25+ costs £999 and the S25 costs £799. The Google Pixel 9 Pro XL costs £799, the OnePlus 13 costs £899 and the Apple iPhone 16 Plus costs £899.

View image in fullscreen Despite being thin and light the S25 Edge is still a big phone and hard to use one-handed without a grip or handle. Photograph: Samuel Gibbs/The Guardian

Verdict

The S25 Edge is an interesting move for Samsung. I don’t think many people have looked at their current phone and wished it could be thinner versus lasting longer or having a better camera. Most people use a case and that immediately nullifies the small differences in thickness of most premium devices.

But using the Edge has reminded me of just how heavy modern big phones can be. It feels so much lighter in the hand, in your pocket or in your bag while still having a big, quality screen. The battery life is certainly not tremendous, but it hasn’t been terrible either, getting through heavy use days with a little left in the tank.

The main camera is excellent but the lack of a good telephoto camera might be a deal-breaker for some. I did miss it. And while it feels very solid with a titanium frame and appears to survive aggressive bending, I would not want to sit on it in a back pocket.

The S25 Edge is a quality piece of hardware, so if you’ve ever wanted a lighter, thinner, big-screen phone, this is it. But for everyone else better options exist from Samsung or others at this price.

Special featherweight, titanium edition of top Android has large screen but sacrifices battery and camera for design

Samsung Galaxy S25 Edge review: super thinness above all else

Self-driving Ubers are expected to appear on roads in London next year after the government said trials of fully autonomous vehicles would be brought forward to spring 2026.

Companies will be allowed to run pilots of small-scale taxi or “bus-like” services for public use – and, for the first time in Europe, without any human safety driver onboard or in the driving seat.

Uber will partner with the UK tech firm Wayve to launch trials of taxis bookable via its app in the capital, its largest European market.

A fuller rollout of self-driving taxis, or robotaxis, will come after the Automated Vehicles Act fully takes effect in late 2027.

The UK has sped up the process now that driverless taxis have become established in San Francisco in the US and numerous cities in China. Uber rolled out its first driverless taxis with the US firm Waymo in Austin, Texas, in March this year, where Tesla is also planning to launch a rival autonomous service this month.

The Department for Transport (DfT) said the technology would make UK roads safer and claimed it could create 38,000 jobs and create an industry worth £42bn by 2035.

The transport secretary, Heidi Alexander, said: “The future of transport is arriving. Self-driving cars could bring jobs, investment, and the opportunity for the UK to be among the world leaders in new technology.

“With road safety at the heart of our pilots and legislation, we continue to take bold steps to create jobs, back British industry, and drive innovation to deliver our plan for change.”

The DfT said self-driving vehicles couldimprove transport for millions of people – adding new public transport options in rural areas and improving accessibility for those unable to drive.

Trials of self-driving vehicles have been taking place in the UK for more than a decade, using technology from Wayve and fellow British company Oxa (formerly Oxbotica). However, so far all road tests of cars or buses have needed a safety driver in the vehicle ready to take over the controls.

The Automated Vehicles Act will require self-driving cars to be approved after tests to demonstrate “a level of safety at least as high as competent and careful human drivers”, the department said.

Previous pledges to bring robotaxis to London have not come to fruition. A driverless bus service launched in Edinburgh in 2023 stopped operating due to a lack of passengers.

However, autonomous taxi services in the US have now operated hundreds of thousands of paid trips. It has not been all plain sailing: General Motors abandoned its plans for an autonomous service after a number of incidents, including one where a taxi dragged and severely injured a pedestrian. But early reports suggest self-driving taxis are safer and some customers, especially women, prefer hiring a driverless taxi.

Alex Kendall, the co-founder and chief executive of Wayve, said the accelerated trials would put the UK in a leading position for fully autonomous cars, adding: “These early pilots will help build public trust and unlock new jobs, services, and markets.”

Gavin Jackson, Oxa’s CEO, said: “Clear rules will open up the market and encourage transport companies to introduce the benefits of autonomous vehicles across the country. Today’s announcement shows that Britain is ready for this technology.”

Partnership with Wayve will use cars without a human safety driver onboard for the first time in Europe

Uber to trial self-driving taxis in London next spring

Artificial intelligence should be deployed to “level up” opportunities for dyslexic children, according to the UK science and technology secretary, Peter Kyle, who warned there was currently not enough human capacity to help people with the learning difficulty.

Kyle, who is dyslexic and uses AI to support his work, said the government should carefully look at “how AI can transform education and help us assess and understand a young person’s abilities into the future”.

He spoke as the TV chef Jamie Oliver, who is also dyslexic, launched a campaign calling for improved teacher training on dyslexia and earlier screening of children to detect the condition sooner. About 6 million people in the UK are estimated to have dyslexia, which primarily affects reading and writing skills.

Kyle told the Guardian he had felt “quite emotional” when seeing AI technology used to help young people learn with “incredible empathy, encouragement and knowledge”.

“I have already seen how GPTs and other [AI] services have helped me,” said Kyle, who was last month mooted in press reports as a potential future education secretary. “AI gets to know you. AI gets to know how you ask questions and how you think. It fits in around your own individual learning characteristics. AI is an incredible tutor, so there is no question that AI deployed wisely and safely, not just in education but in a young person’s life, can have an incredible levelling-up opportunity.”

He said the current problem was that “we don’t have enough human capacity to give dyslexics all of the skilled and specialist support that is unique to the individual characteristics of dyslexics”.

He stopped short of arguing AI technology should be allowed to support dyslexic children in exams, where they perform significantly worse in key GCSEs. About 52% of children without any special education needs achieved a grade 5 or above in English and maths last year, compared with 22% among children with a specific learning difficulty, which includes dyslexia.

“We are currently reviewing the curriculum and we have to very carefully look at how AI can transform education and help us assess and understand a young person’s abilities into the future,” he said. “But right now I think we have a really robust exam system. It’s very good at judging a young person’s potential.”

Kay Carter, the chief executive of the Dyslexia Association, said AI was already levelling the playing field for dyslexic pupils. If AI could manage tasks such as memorising facts and rapid recall of information, “the focus [of education] may shift to problem-solving [and] critical thinking, talents which some of those with dyslexia naturally excel at,” she said.She cautioned that AI was not to be a replacement for good teaching but “allows dyslexic students better access to their own learning”.

Kyle was speaking at London Tech Week, where he also addressed the row between the government and parts of the creative industries over the use of copyrighted content for training AI models. Elton John last month called Kyle “a bit of a moron” after ministers pushed back against a campaign for the new data bill to provide greater protections for creatives.

skip past newsletter promotion Sign up to First Edition Free daily newsletter Our morning email breaks down the key stories of the day, telling you what’s happening and why it matters Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

“I am always available to meet with Elton John and anyone else,” Kyle said. “I have met with Björn [Ulvaeus] from Abba and publishers, I have spoken to small creatives.”

Kyle said the data bill, which is currently rallying back and forth between MPs and peers, was “totally not suitable” to legislate on AI using copyrighted material. The Guardian reported last week that the government has decided to introduce a “comprehensive” AI bill in the next parliamentary session to address concerns about issues including safety and copyright.

“I will set up working groups the very second the data bill is through parliament so I can begin the rapid process towards legislation,” he said.

Peter Kyle, who is dyslexic and uses AI in his work, says government should look at how it ‘can transform education’

AI can ‘level up’ opportunities for dyslexic children, says UK tech secretary

Britain’s media regulator, Ofcom, on Tuesday launched nine investigations into the internet message board 4chan as well as several file-sharing services over possible breaches of online safety laws.

Britain’s Online Safety Act, passed in 2023, sets tougher standards for platforms to tackle criminal activity, with an emphasis on child protection and illegal content.

The regulator said it had received complaints about potential illegal content on 4chan and, separately, possible sharing of child sexual abuse material on seven file-sharing services.

Ofcom is investigating whether 4chan and the file-sharing services failed to put in place appropriate safety measures to protect users from illegal content and to respond to a statutory information request. It also wants to assess whether they kept proper risk assessment records. 4chan could not be reached for comment.

The regulator can require platforms to take specific measures to ensure compliance with the rules, or fine them up to £18m ($24.3m) or 10% of qualifying worldwide revenue, whichever is greater.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Ofcom’s announcement also said it was investigating whether First Time Videos, which provides pornographic services, has effective enough age assurance in place to protect children.

Ofcom looks into whether 4chan and file-sharing services failed to put measures to protect users from illegal content

UK media regulator investigates possible online safety breaches at 4chan

In the movie “Jurassic Park,” a character proclaims there is nothing to fear from a towering Brachiosaurus, because it’s a “veggie-saurus” that eats only plants. Littlefoot, the “Longneck” dinosaur in the “Land Before Time” series, chows down on leaves, or “tree stars.” But while pop culture and general scientific opinion have agreed for decades that the long-necked sauropod dinosaurs were herbivores, there was no definitive proof found in the fossil record.

But there were hints of a diet full of green stuff.

Fossils of sauropods, which stomped across the planet for 130 million years, are plentiful; additionally, herbivores tend to outnumber those of carnivores. The animals had small, peg-like teeth, and their huge, lumbering bodies seemed ill-equipped to chase down prey. “Plants were pretty much the only option,” said Stephen Poropat, a paleontologist and the deputy director of the Western Australian Organic and Isotope Geochemistry Center at Curtin University in Perth.

A study published Monday in the journal Current Biology provides what may be the first concrete proof to support this argument, in the form of fossilized plants discovered in the belly of a sauropod. “It’s the smoking gun, or the steaming guts, as it were — the actual direct evidence in the belly of the beast,” Dr. Poropat said. “It’s never been found before for a sauropod dinosaur.”

“It’s the smoking gun, or the steaming guts,” said a paleontologist from a team that also found that the sauropods didn’t chew their food.

First Fossil Proof Found That Long-Necked Dinosaurs Were Vegetarians

The US administration’s targeting of academic research and international students is a “great gift” to China in the race to compete on artificial intelligence, former OpenAI board member Helen Toner has said.

The director of strategy at Georgetown’s Center for Security and Emerging Technology (CSET) joined the board of OpenAI in 2021 after a career studying AI and the relationship between the United States and China.

Toner, a 33-year-old University of Melbourne graduate, was on the board for two years until a falling out with founder Sam Altman in 2023. Altman was fired by the board over claims that he was not “consistently candid” in his communications and the board did not have confidence in Altman’s ability to lead.

The chaotic months that followed saw Altman fired and then re-hired with three members of the board, including Toner, ousted instead. They will soon also be the subject of a planned film, with the director of Challengers and Call Me By Your Name, Luca Guadagnino, reportedly in talks to direct.

The saga, according to Time magazine – which named her one of the Top 100 most influential people on AI in 2024 – resulted in the Australian having “the ear of policymakers around the world trying to regulate AI”.

At CSET, Toner has a team of 60 people working on AI research for white papers or briefing policymakers focused on the use of AI in the military, workforce, biosecurity and cybersecurity sectors.

“A lot of my work focuses on some combination of AI, safety and security issues, the Chinese AI ecosystem and also what gets called frontier AI,” Toner said.

Toner said the United States is concerned about losing the AI race to China and while US chip export controls make it harder for China to get compute power to compete with the US, the country was still making a “serious push” on AI, as highlighted by the surprise success of Chinese generative AI model DeepSeek earlier this year.

The Trump administration’s attacks on research and bans on international students are a “gift” to China in the AI race with the US, Toner said.

“Certainly it’s a great gift to [China] the way that the US is currently attacking scientific research, and foreign talent – which is a huge proportion of the USA workforce – is immigrants, many of them coming from China,” she said.

“That is a big … boon to China in terms of competing with the US.”

The AI boom has led to claims and concerns about a job wipeout caused by companies using AI to replace work that had otherwise been done by humans. Dario Amodei, the CEO of Anthropic, the company behind the generative AI model Claude, told Axios last week that AI could reduce entry-level white-collar jobs by 50% and result in 20% unemployment in the next five years.

Toner said Amodei “often says things that seem directionally right to me, but in terms of … timeline and numbers often seem quite aggressive” but added that disruption in the jobs market had already started to show.

“The kind of things that [language model-based AI] can do best at the moment … if you can give them a bite-size task – not a really long term project, but something that you might not need ages and ages to do and something where you still need human review,” she said. “That’s a lot of the sort of work that you give to interns or new grads in white-collar industries.”

Experts have suggested companies that invested heavily in AI are now being pressed to show the results of that investment. Toner said while the real-world use of AI can generate a lot of value, it is less clear what business models and which players will benefit from that value.

Dominant uses might be a mix of different AI services plugged into existing applications – like phone keyboards that can now transcribe voices – as well as stand-alone chatbots, but it’s “up in the air” which type of AI would actually dominate, she said.

skip past newsletter promotion Sign up to Breaking News Australia Free newsletter Get the most important news as it breaks Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

Turner said the push for profitability was less risky than the overall race to be first in AI advancements.

“It means that these companies are all making it up as they go along and figuring out as they go how to make trade-offs between getting products out the door, doing extra testing, putting in extra guardrails, putting in measures that are supposed to make the model more safe but also make it more annoying to use,” she said.

“They’re figuring that all out on the fly, and … they’re making those decisions while under pressure to go as fast as they can.”

Toner said she was worried about the idea of “gradual disempowerment to AI” – “meaning a world where we just gradually hand over more control over different parts of society and the economy and government to AI systems, and then realise a bit too late that it’s not going the way that we wanted, but we can’t really turn back”.

She is most optimistic about AI’s use in improving science and drug discovery and for self-driving services like Waymo in reducing fatalities on the roads.

“With AI, you never want to be looking for making the AI perfect, you want it to be better than the alternative. And when it comes to cars, the alternative is thousands of people dying per year.

“If you can improve on that, that’s amazing. You’re saving many, many people.”

Toner joked that her friends had been sending her options on who might play her in the film.

“Any of the names that friends of mine have thrown my way are all these incredibly beautiful actresses,” she said. “So I’ll take any of those, whoever they choose.”

Influential researcher claims disruption in jobs market from generative AI has already begun and warns of possibility of ‘gradual disempowerment to AI’

US attacks on science and research a ‘great gift’ to China on artificial intelligence, former OpenAI board member says

Skincare regimes demonstrated by young influencers on TikTok offer little to no benefit, researchers have found, adding that on the contrary they raise the risk of skin irritations and lifelong allergies in children.

The team behind the study say there has been a rise in young girls sharing videos of complex skincare routines with moisturisers, toners, acne treatments and anti-ageing products.

In the first study of its kind, researchers analysed such videos on the plaftorm and found the regimes are not only laborious – some girls wake up as early as 4.30am to fit in their routines – but pricey, unnecessary and potentially harmful.

“The risks associated with using these products, especially in young girls, far outweighs whatever marginal benefit you may get from using the active ingredients,” said Dr Molly Hales, the first author of the research from Northwestern University.

Hales added that while social media can be a place for playfulness and self-expression, young girls do not need such regimes.

“The danger is when girls get the message that this is something that they have to do to take care of their skin and to protect their health,” she said.

“These products don’t increase the health of your skin and they probably worsen your skin integrity over time.”

Instead she said a gentle cleanser once or twice a day and the application of sunscreen is sufficient.

Writing in the journal Paediatrics, Hales and colleagues report how they created two TikTok accounts, purportedly for 13-year-olds and, and accumulated a sample of 100 skincare routines videos filmed by young people.

The team found all but one of the 82 creators were girls, and their age ranged from seven to 18 years old.

“Almost all content creators had clear, light skin without visible blemishes,” the researchers write, adding the videos “made frequent reference to aspirational beauty ideals that may be tied to whiteness”.

The study reveals the skincare regimes involved six products on average, often from the same brands, with a total average cost of $168 (£124). However, some involved more than a dozen products with a total cost of more than $500 (£369).

“Get Ready With Me” videos were the most common, followed by “Skin Care Routine” and “After School” skincare routine videos.

The team found many of the ingredients in the featured products, such as citric acid, carry a risk of skin irritation and sun sensitivity – meaning they increase the chance of sunburn and other skin damage. Yet only 26% of the 84 videos showing daytime skincare routines included a sunscreen – something that authors called “a significant missed opportunity”.

Hales added the use of products with multiple active ingredients, or layering products with the same active ingredients, increases the risk of skin irritation.

The researchers found that among the 25 top-viewed videos, 76% contained at least one potential contact allergen – often fragrance.

Prof Tess McPherson of the British Association of Dermatologists, who was not involved in the work, said the study was important, backing up anecdotal reports of an increase in young people attending clinics with skin irritation.

“We’re certainly seeing a huge increase in people wanting something they call perfect or flawless skin, which we know is unachievable [and] unhelpful – it doesn’t make people happy,” she said.

McPherson added she is aware of young people asking for birthday money to put towards expensive skincare products, many of which are now being targeted at young girls through appealing packaging.

As well as the risks highlighted by the study, McPherson said the skincare videos contribute to an increase in stigma around conditions such as acne and eczema as well as a fear of ageing, a message she said was reinforced by women embracing Botox and fillers.

“Younger and younger children are seeking skincare products when they don’t need them, they’re not helpful,” she said. “This is a very concerning statement on society and how we view how skin should look.”

With number of young girls sharing videos rising, study says following instructions can irritate skin and lead to allergies

Teenage girls’ TikTok skincare regimes offer little to no benefit, research shows

Artificial intelligence companies are plundering 130,000 film and TV scripts to train their models in a raid on copyrighted material that poses a “direct threat” to the future of the UK screen sector, according to the British Film Institute.

In a wide-ranging report analysing the benefits and threats posed by AI to the UK’s film, TV, video game and visual special effects industries, the BFI also raises fears that automation will eliminate the entry-level jobs that bring in the next generation of workers.

It says the “primary issue” facing the £125bn industry is the use of intellectual property (IP) to train generative AI models without payment to, or permission from, rights holders.

The UK creative industries want to see an “opt-in” regime, forcing AI companies to seek permission and strike licensing deals before they can use content, and the government is currently in the process of considering what legislation to put in place.

“AI offers significant opportunities for the screen sector such as speeding up production workflows, democratising content creation and empowering new voices,” said Rishi Coupland, director of research and innovation at the BFI. “However, it could also erode traditional business models, displace skilled workers, and undermine public trust in screen content.”

While the report acknowledges the technological benefits of AI – such as being able to de-age actors and improving the authenticity of accents, as was used controversially in Adrien Brody’s Oscar-winning film The Brutalist – it also cites fears of job losses.

Tasks such as writing, translation and some technical visual effects and character animation can now be automated, “prompting fears of obsolescence among professionals” amid concerns that AI could take over the jobs once done by the sector’s youngest workers.

“AI’s ability to automate tasks raises fears of job losses, particularly for junior or entry-level positions,” the report says. “Training and upskilling are seen as essential to prepare the workforce for AI integration.”

However, the report, published in partnership with Goldsmiths, Loughborough and Edinburgh universities, warns of a “critical shortfall” in AI training provision.

“AI education in the UK screen sector is currently more ‘informal’ than ‘formal’,” the report says. “And many workers – particularly freelancers – lack access to resources that would support them to develop skills complementary to AI.”

More than 13,000 creative technology companies are based in the UK, including more than 4,000 businesses focused on applying emerging technologies across film, games and other creative subsectors.

London is the world’s second largest hub after Mumbai for visual effects professionals globally, home to leading businesses including Framestore, whose credits include Avengers: Endgame and the BBC’s adaptation of His Dark Materials.

The report also acknowledges that AI is likely to significantly benefit the industry by lowering barriers for creators “regardless of budget or experience”.

“This could empower a new wave of British creators to produce high-quality content with modest resources,” it says.

skip past newsletter promotion Sign up to Business Today Free daily newsletter Get set for the working day – we'll point you to all the business news and analysis you need every morning Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The 45-page report makes nine recommendations including establishing a market for IP licensing and training, the development of “market-preferred, culturally inclusive AI tools” and more investment in the sector and skills training.

Quick Guide Contact us about this story Show The best public interest journalism relies on first-hand accounts from people in the know. If you have something to share on this subject you can contact us confidentially using the following methods. Secure Messaging in the Guardian app The Guardian app has a tool to send tips about stories. Messages are end to end encrypted and concealed within the routine activity that every Guardian mobile app performs. This prevents an observer from knowing that you are communicating with us at all, let alone what is being said. If you don't already have the Guardian app, download it (iOS/Android) and go to the menu. Select 'Secure Messaging'. SecureDrop, instant messengers, email, telephone and post See our guide at theguardian.com/tips for alternative methods and the pros and cons of each. Illustration: Guardian Design / Rich Cousins Was this helpful? Thank you for your feedback.

The report was also produced in conjunction with the CoStar Foresight Lab, the £75.6m national network of laboratories that is developing new technology for the industry.

“AI offers powerful tools to enhance creativity, efficiency, and competitiveness across every stage of the production workflow,” said Jonny Freeman, director of CoStar. “From script development and pre-production planning, through on-set production, to post-production and distribution. However, it also raises urgent questions around skills, workforce adaptation, ethics and sector sustainability.”

Last week, the BBC director general and the boss of Sky criticised proposals to let tech firms use copyright-protected work without permission, while Lisa Nandy, the culture secretary, tried to reassure the creative industry that it would not be harmed by future AI legislation.

“We approach you with no preferred option in mind,” she said, delivering a keynote speech at the Deloitte Enders conference. “We are a Labour government, and the principle [that] people must be paid for their work is foundational. You have our word that if it doesn’t work for the creative industries, it will not work for us.”

Film institute’s report raises fears AI will eliminate junior roles in film, TV, video game and special effects industries

AI plundering scripts poses ‘direct threat’ to UK screen sector, says BFI

Bruno Pereira has been considered one of the great Indigenous protectors of his generation. And this has made him an enemy of a man called Amarildo da Costa de Oliveira, also known as Pelado. The Guardian’s Latin America correspondent Tom Phillips reports on the story of these two men – Bruno and Pelado – and what happened when their paths collide

Episode three: the protector and the poacher

From using motion capture tech to allow the Indian cricketing star Rahul Dravid to give personalised coaching tips for children to an algorithm trained on Shakespeare’s handwriting powering a robotic arm to rewrite Romeo and Juliet, artificial intelligence is rapidly revolutionising the global advertising industry.

Those AI-created adverts, for the Cadbury’s drink brand Bournvita and the pen maker Bic, were produced by agency group WPP, which is spending £300m annually on data, tech and machine learning to remain competitive.

Mark Read, the chief executive of the London-listed marketing services group, has said AI is “fundamental” to the future of its business, while admitting that it will drastically reshape the ad industry workforce.

Now Read has announced he is to leave at the end of this year, after almost seven years as chief executive and more than 30 at WPP, as the company struggles to keep pace with its peers and also counter moves by big tech to muscle in to the AI-driven future of advertising.

For ad agencies, the upheaval originates from a familiar source. Over more than a decade, Google and the Facebook owner, Meta, successfully built tech tools for publishers and ad buyers that helped them to dominate online. Big tech hoovered up almost two-thirds of the £45bn spent by advertisers in the UK this year. Now, Mark Zuckerberg wants to take over making the ads, too.

The WPP firm VML used AI to create the ‘One Bic, One Book, Two Classics’ campaign for Bic in Brazil. Photograph: WPP

The Meta boss is gearing up to unleash AI tools to allow advertisers to fully create and target campaigns on his social media sites, prompting fears of the “death of creativity” – and widespread job cuts at agencies.

Last week it emerged that these tools are to be rolled out by the end of next year, with Zuckerberg describing the capability in a recent interview as a “redefinition of the category of advertising. You don’t need any creative, you don’t need any targeting, you don’t need any measurement, except to be able to read the results that we spit out,” he said last month, in comments that appear to render much of the advertising industry obsolete.

Rahul Dravid in Cadbury’s Bournvita campaign, which used AI.

Agencies of all sizes – and particularly the deep-pocketed international groups such as WPP, Publicis and Omnicom – are pouring investment into developing their own AI tools and working with tech companies such as Meta and Google. But the plan is meant to be to keep clients, not lose them.

“I think there is no doubt AI will disintermediate a large number of jobs,” says the chief executive of one big ad agency. “Having said that, there are many agencies with big corporate clients that really could do with being slimmed down a lot. I can see staffing in areas like strategy, consumer insight and some conceptual roles being safe, but what will be really hit is those involved in production and the realisation of ideas.”

Big tech executives espoused the benefits of AI at the annual Enders Deloitte conference on the media and telecoms industry last week.

Stephan Pretorius, who described himself as “WPP’s AI guy” as he co-led a session with Meta, said: “Creativity, in its purest form, remains a human skill.”

WPP’s Stephan Pretorius on AI and advertising.

He argued that AI does not equate to job cuts but admitted that agencies need to restructure and advertising client relationships are changing.

“AI replaces tasks, it eliminates tasks, it doesn’t eliminate jobs,” he said. “A lot of what we used to get paid for is now getting automated and therefore our commercial models have to change, our team structures have to change. The way we are incentivised by our clients is changing. But that is the transition.”

Last month WPP said there would be an undisclosed number of redundancies globally across WPP Media, formerly known as GroupM, which plans and buys billions of pounds of ad space for clients across digital and traditional media.

“You have a situation where the big holding companies are in a dilemma,” another ad agency chief executive says. “Clients expect us to invest millions developing AI so they can cut their budgets because things can be done quicker and cheaper. Lots of clients are asking for fee reductions.”

So far the AI revolution does not appear to be having a big impact on the UK industry.

View image in fullscreen The Facebook and Instagram owner, Meta, has said it will release AI tools to allow advertisers to fully create and target campaigns on social media sites. Photograph: Anadolu/Getty Images

There were a record 26,787 people employed in media, creative and digital agencies last year, according to the Institute of Practitioners in Advertising (IPA), the trade body for agencies that represent 85% of ad spend by clients in the UK.

skip past newsletter promotion Sign up to Business Today Free daily newsletter Get set for the working day – we'll point you to all the business news and analysis you need every morning Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The IPA has been measuring the size of the market since 1960, when there were 19,000 employees, hitting a low point of just under 12,000 in the early 1990s.

The amount spent on advertising has also grown exponentially, driven by the advent of the internet era, from £60m recorded in the pre-TV era in 1938.

By 1982 the UK market was worth £3.1bn and this year it is forecast to cross £45bn, according to the Advertising Association/Warc, which has been publishing figures annually since 1980.

Agency bosses believe that for the biggest household name advertisers there is too much brand risk handing over the full creative process to AI, which, for now at least, does not have the capability to make top-quality ads.

“You can often tell a [pure] AI piece of work a thousand yards away – glossy, very idealised and slightly plasticky looking,” the chief of one creative agency says. “But that will change. You hear creatives saying that AI is never going to come up with something as brilliant as, say, a gorilla playing drums for Cadbury, but I am not so sure. AI will eventually become fine-tuned enough to react to the quite left-field conceptual prompting.”

View image in fullscreen The Cadbury’s Dairy Milk TV advert featuring a gorilla playing the drums – created by the ad agency Fallon – became a viral hit. Photograph: Rex Features

Since making the comments that the industry took to mean that Meta is seeking to supplant the role of agencies, Zuckerberg has tried to clarify that the AI tools are primarily expected to be used by small and medium-sized businesses.

“In the future, if you were working with a creative agency to make creative, you’ll probably keep doing that,” he said at the Stripe conference, clarifying his position a week after making his initial comments about the scope of Meta’s AI advertising plans. “If you aren’t and you’re just hacking something together and throwing it into Meta’s ad system, well now we’re going to be able to come up with 4,000 different versions of your creative and just test them and figure out which one works best.”

Meta and Google have always considered that they have “democratised” advertising by enabling the long tail of millions of small businesses that do not have the financial wherewithal to run TV ad campaigns, or employ an ad agency to run campaigns.

“That is the smokescreen they always use,” the ad agency boss says. “When they first emerged as new ad channels decades ago it was all about small businesses, and now they take nearly two-thirds of all UK ad spend.”

In the noughties, as big tech grew increasingly powerful, Sir Martin Sorrell, who built WPP into the world’s biggest ad group and is now the chief executive of S4 Capital, labelled Meta and Google “frenemies” – meaning they can be seen as a partner and as a competitive threat to agencies.

Two decades on, the rise of AI in advertising is the latest technological development forcing the industry to once again adapt to survive.

“Meta’s new promise to ‘auto-generate your ad in seconds’ is the clearest sign yet that the production sausage factory is about to be fully mechanised,” says Patrick Garvey, the co-founder of the independent agency We Are Pi. “It’s not the death of agencies. It’s the death of outdated agency models.”

He is supportive of small businesses benefiting from the changes but says Meta’s approach to AI is akin to the “fast food of advertising”. For traditional companies in adland, it could be a difficult meal to stomach.

WPP and others roll out AI-generated campaigns as Facebook owner Meta plans to let firms create their own ads

The ‘death of creativity’? AI job fears stalk advertising industry

All civil servants in England and Wales will get practical training in how to use artificial intelligence to speed up their work from this autumn, the Guardian has learned.

More than 400,000 civil servants will be informed of the training on Monday afternoon, which is part of a drive by the chancellor of the duchy of Lancaster, Pat McFadden, to overhaul the civil service and improve its productivity.

At the same time, the size of the civil service is being reduced by tens of thousands of roles through voluntary redundancy and not replacing leavers. The government said officials would be tasked with figuring how they could use AI technology to streamline their own work wherever possible.

Officials are already piloting a package of AI tools called Humphrey – named after the senior civil servant Sir Humphrey Appleby from the 1980s TV sitcom Yes, Minister.

In March, it was announced that officials would be told to abide by a mantra that says: “No person’s substantive time should be spent on a task where digital or AI can do it better, quicker and to the same high quality and standard.”

The practical training, set to roll out this autumn, will give all civil servants a working-level knowledge of AI and show how it is already being used to cut the time spent on everyday tasks.

A recent pilot of AI by the Scottish government used a Humphrey tool to sift consultation responses about regulating cosmetic surgery such as lip filler procedures and found its results were similar to those produced by humans but it was quicker.

No decision has yet been taken on its wider rollout but the government said across 500 annual consultations the tool could help spare officials from as much as 75,000 days of analysis every year, which costs the government £20m in staffing costs.

The Department for Work and Pensions is also using AI to understand high volumes of correspondence, which used to take weeks. This allows benefits or pensions recipients to be identified as potentially vulnerable and in need of urgent support.

In a message to all staff, Sir Chris Wormald, the cabinet secretary, said he was “proud” of the civil service’s “ability to continuously adapt to new challenges” as he urged them to consider how it “needs to evolve and reform”.

skip past newsletter promotion Sign up to First Edition Free daily newsletter Our morning email breaks down the key stories of the day, telling you what’s happening and why it matters Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

He said: “The prime minister has set us an important task in building a productive and agile state, which will involve us preserving and championing everything that is great about the civil service while changing to meet the challenges of an uncertain world.”

He added: “We need to seize the opportunities of artificial intelligence and other technological developments to continually modernise everything that we do.”

The training will be provided through the One Big Thing project, which focuses on a different key skill each year.

While the government is rapidly embracing AI in Whitehall, some are sounding the alarm about its potential to introduce or reinforce errors and bias without adequate human oversight. The Guardian revealed in December that an artificial intelligence system used by the UK government to detect welfare fraud had been showing bias according to people’s age, disability, marital status and nationality.

Officials are piloting package of AI tools called Humphrey – named after character in TV sitcom Yes, Minister

All civil servants in England and Wales to get AI training

Bereaved parents saw their hopes for change dashed after a bill meant to protect children from sexual predators and drug dealers online died in the Colorado state legislature last month.

Several of those parents had helped shape the bill, including Lori Schott, whose 18-year-old daughter Annalee died by suicide in 2020 after consuming content on TikTok and Instagram about depression, anxiety and suicide.

“When the legislators failed to vote and pushed it off onto some fake calendar date where they’re not even in session, to not even have accountability for where they stand – as a parent, it’s a slap in the face,” said Schott, who identifies as a pro-second amendment Republican. “It’s a slap in the face of my daughter, and to other kids that we’ve lost.”

Had the legislation passed, it would have required social media platforms like Facebook, Instagram and TikTok to investigate and take down accounts engaged in gun or drug sales or in the sexual exploitation or trafficking of minors. It also mandated the creation of direct hotlines to tech company personnel for law enforcement and a 72-hour response window for police requests, a higher burden than under current law.

Additionally, platforms would have had to report on how many minors used their services, how often they did so, for how long and how much those young users engaged with content that violated company policies. Several big tech firms registered official positions on the bill. According to Colorado lobbying disclosures, Meta’s longtime in-state lobby firm, Headwater Strategies, is registered as a proponent for changing the bill. Google and TikTok also hired lobbyists to oppose it.

View image in fullscreen ‘[Legislators] chose big tech over protecting children and families.’ Illustration: Andrei Cojocaru/Guardian

“We’re just extremely disappointed,” said Kim Osterman, whose 18-year-old son Max died in 2021 after purchasing drugs spiked with fentanyl from a dealer he met on Snapchat. “[Legislators] chose big tech over protecting children and families.”

Colorado legislators agree to hold social media companies responsible for protecting children

Protections for users of social media (SB 25-086) passed both chambers before being vetoed on 24 April by governor Jared Polis, a Democrat, who cited the bill’s potential to “erode privacy, freedom and innovation” as reasons for his veto. Colorado’s senate voted to override the veto on 25 April, yet those efforts fell apart on 28 April when the state house opted to delay the vote until after the legislative session ended, effectively blocking an override and keeping the bill alive.

The bill originally passed the senate by a 29-6 vote and the house by a 46-18 margin. On 25 April, the senate voted 29-6 to override Polis’s veto. Lawmakers anticipated that the house would take up the override later that day. At the time, according to those interviewed, there appeared to be enough bipartisan support to successfully overturn his veto.

“It was an easy vote for folks because of what we were voting on: protecting kids from social media companies,” said the senator Lindsey Daugherty, a Democrat and a co-sponsor of the bill. She said she urged house leadership to hold the vote Friday, but they declined: “The speaker knew the governor didn’t want us to do it on Friday, because they knew we would win.”

The parents who advocated for the bill attribute its failure to an unexpected, 11th-hour lobbying campaign by a far-right gun owners’ association in Colorado. Two state legislators as well as seven people involved in the legislative process echoed the parents’ claims.

An abnormal, last-minute campaign disrupts bipartisan consensus

Rocky Mountain Gun Owners (RMGO) cast the bill as an instrument of government censorship in texts and emails over the legislation’s provisions against “ghost guns”, untraceable weapons assembled from kits purchased online, which would have been prohibited.

RMGO launched massive social media and email campaigns urging its 200,000 members to contact their legislators to demand they vote against the bill. A source with knowledge of the workings of the Colorado state house described the gun group’s social media and text campaigns, encouraging Republicans voters to contact their legislators to demand opposition to the bill, as incessant.

“[Legislators] were getting countless calls and emails and being yelled at by activists. It was a full-fledged attack. There was a whole campaign saying: ‘This is a government censorship bill,’” they said.

The group’s actions were instrumental in a campaign to deter house Republicans from voting against the veto, resulting in the quashing of the bill, and unexpected from an organization that had been facing funding shortfalls, according to 10 people interviewed who were involved in the design of the bill and legislative process. Sources in the Colorado state house spoke to the Guardian on condition of anonymity out of fear of reprisal from RMGO.

The house had delayed the vote until 28 April, which allowed RMGO time to launch a campaign against the bill over the weekend. When lawmakers reconvened Monday, the house voted 51-13 to postpone the override until after the legislative session ended – effectively killing the effort.

View image in fullscreen ‘It was a full-fledged attack. There was a whole campaign saying: “This is a government censorship bill.”’ Illustration: Andrei Cojocaru/Guardian

The gun activists’ mass text message campaign to registered Republican voters asserted the social media bill would constitute an attempt to “compel social media companies to conduct mass surveillance of content posted on their platforms” to search for violations of Colorado’s gun laws, describing the bill as an attack on first and second amendment rights, according to texts seen by the Guardian.

A familiar, aggressive foe

Founded in 1996, RMGO claims to have a membership of more than 200,000 activists. It is recognized as a far-right group that takes a “no-compromise” stance on gun rights. Dudley Brown, its founder and leader, also serves as the president of the National Association for Gun Rights, which positions itself further to the right than the National Rifle Association (NRA). RMGO has mounted criticism against the NRA for being too moderate and politically compromising. Critics have described RMGO as “bullies” and “extremists” because of its combative tactics, which include targeting and smearing Democrats and moderate Republicans. The group did not respond to requests for comment on its legislative efforts.

RMGO is a well-known presence at the Colorado capitol, typically opposing gun-control legislation. Daugherty described its typical campaign tactics as “scary”. She got rid of her X account after being singled out by the group over her work on a bill to ban assault weapons earlier this year.

“When we were running any of the gun bills at the capitol, they put my and some other legislators’ faces on their websites,” she said. A screenshot of a tweet from RMGO showed Daugherty with a red “traitor” stamp on her forehead.

The group’s campaign resulted in the spread of misinformation about the bill’s impact on gun ownership rights, sources involved in the legislative process said.

“The reason I was in support of the bill, and in support of the override, was it has to do with child trafficking and protecting the kids,” said the senator Rod Pelton, a Republican, who voted in favor of the veto override in the senate. “I just didn’t really buy into the whole second amendment argument.”

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The bill had enjoyed the backing of all 23 of Colorado’s district attorneys as well as bipartisan state house support.

RMGO’s late-stage opposition to the social media bill marked a break from its usual playbook. The group generally weighs in on legislation earlier in the process, according to eight sources, including two of the bill’s co-sponsors, Daugherty and the representative Andy Boesenecker.

“They really ramped up their efforts,” Boesenecker said. “It was curious to me that their opposition came in very late and appeared to be very well funded at the end.”

In recent years, RMGO group had been less active due to well-documented money problems that limited its ability to campaign on legislative issues. In a 2024 interview, the group’s leaders stated plainly that it struggled with funding. Daugherty believes RMGO would not have been able to embark on such an apparently costly outreach campaign without a major infusion of cash. A major text campaign like the one launched for SB-86 was beyond their financial capacity, she said. Others in Colorado politics agreed.

“Rocky Mountain Gun Owners have not been important or effective in probably at least four years in the legislature. They’ve had no money, and then all of a sudden they had tons of money, funding their rise back into power,” said Dawn Reinfeld, executive director of Blue Rising Together, a Colorado-based non-profit focused on youth rights.

The campaign made legislators feel threatened, with primary elections in their districts over the weekend, Daugherty said, particularly after accounts on X, formerly Twitter, bombarded the bill’s supporters.

View image in fullscreen ‘The bill gave me hope that Avery’s legacy would be to help. So when it didn’t pass, it was pretty soul-crushing.’ Illustration: Andrei Cojocaru/Guardian

“Folks were worried about being primaried, mostly the Republicans, and that’s kind of what it came down to,” Daugherty said.

Aaron Ping’s 16-year-old son Avery died of an overdose in December after buying what he thought was ecstasy over Snapchat and receiving instead a substance laced with fentanyl. Ping saw the campaign against the bill as an intentional misconstrual of its intent.

“It was looking like the bill was going to pass, until all this misinformation about it taking away people’s gun rights because it addresses people buying illegal shadow guns off the internet,” he said.

Ping gave testimony in support of the bill in February before the first senate vote, alongside other bereaved parents, teens in recovery and a district attorney.

“The bill gave me hope that Avery’s legacy would be to help. So when it didn’t pass, it was pretty soul-crushing,” said Ping.

States take up online child-safety bills as federal lawmakers falter

Several states, including California, Maryland, Vermont, Minnesota, Hawaii, Illinois, New Mexico, South Carolina and Nevada, have introduced legislation aimed at improving online safety for children in the past two years. These efforts have faced strong resistance from the tech industry, including heavy lobbying and lawsuits.

Maryland became the first state to successfully pass a Kids Code bill, signing it into law in May 2024. But the victory may be short-lived: NetChoice, a tech industry coalition representing companies including Meta, Google and Amazon, quickly launched a legal challenge against the measure, which is ongoing.

Meanwhile, in the US federal government, the kids online safety act (Kosa), which had wound its way through the legislature for years, died in February when it failed to pass in the House after years of markups and votes. A revamped version of the bill was reintroduced to Congress on 14 May.

In California, a similar bill known as the age-appropriate design code act, modeled after UK legislation, was blocked in late 2023. A federal judge granted NetChoice a preliminary injunction, citing potential violations of the first amendment, which stopped the law from going into effect.

Parents who lost children to online harms helped draft the bill – then watched it collapse under political pressure

They hoped their children’s deaths would bring change. Then a Colorado bill to protect kids online failed

A London-based artificial intelligence company, Stability AI, has claimed that a copyright case brought by the global photography agency Getty Images represents an “overt threat” to the generative AI industry.

Getty’s case against Stability AI for copyright and trademark infringement relating to its vast photography archives reached the high court in London on Monday.

Stability allows users to generate images using text prompts, and its directors include James Cameron, the Oscar-winning film director of Avatar and Titanic. But Getty called the people who were training the AI system “a bunch of tech geeks” and claimed they were indifferent to the problems their innovation might create.

Stability countered by alleging that Getty was using “fanciful” legal routes and spending approximately £10m to fight a technology it feared was “an existential threat” to its business.

Getty syndicates the work of about 50,000 photographers to customers in more than 200 countries. It alleges Stability trained its image generation model on its vast database of copyrighted photographs. As a result the program, called Stability Diffusion, outputs images with Getty Images watermarks still on them. Getty alleges that Stability was “completely indifferent to what they fed into the training data”. It told the court the system amounted to “sticking our trademark on pornography” and “AI rubbish”.

Lawyers for Getty said the dispute over the unlicensed use of thousands of its photographs, including images of celebrities, politicians and news events, “is not a battle between creatives and technology, where a win for Getty Images means the end of AI”.

They added: “The problem is when AI companies such as Stability want to use those works without payment.”

Lindsay Lane KC, representing Getty Images, said: “This was a bunch of tech geeks who were so excited by AI that they were indifferent to any of the dangers or problems it presents.”

In submissions to the court on Monday Getty claimed Stability had trained its image generation model on databases that contained child sexual abuse material.

Stability is fighting the overall Getty claim and its lawyer said the allegations relating to child sexual abuse material were “repugnant”.

A spokesperson for Stability AI said it was committed to preventing misuse of its technology, “particularly in the creation and dissemination of harmful content, including CSAM [child sexual abuse material]”. It said it had robust safeguards “to enhance our safety standards to protect against bad actors”.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The case comes amid a wider campaign from artists, writers and musicians, including Elton John and Dua Lipa, to protect their copyright from alleged theft by generative AI companies, which then use it to allow their customers to create new pictures, music or text.

The UK parliament is locked in a similar dispute after the government proposed that copyright holders would have to opt out of their material being used to train algorithms and produce AI-generated content, otherwise it would be free to use by tech companies.

“Getty Images, of course, recognises that the AI industry overall may be a force for good, but that doesn’t justify allowing those developing AI models to ride roughshod over intellectual property rights,” said Lane.

The trial, which is scheduled to run for several weeks, will focus in part on the use of images by celebrated photographers, including photos of the former Liverpool football coach Jürgen Klopp taken by the award-winning British sports photographer Andrew Livesey; a picture of the Chicago Cubs baseball team by Gregory Shamus, an American sports photographer; a photo of the actor and musician Donald Glover by Alberto Rodriguez; and photos of the actor Eric Dane and film director Christopher Nolan by Andreas Rentz.

Seventy-eight thousand pages of evidence have been disclosed in the case and AI experts are being called to give evidence from the University of California, Berkeley and the University of Freiberg in Germany.

Photography agency alleges Stability AI trained its image generation model on archive of copyrighted pictures

London AI firm says Getty copyright case poses ‘overt threat’ to industry

Apple researchers have found “fundamental limitations” in cutting-edge artificial intelligence models, in a paper raising doubts about the technology industry’s race to develop ever more powerful systems.

Apple said in a paper published at the weekend that large reasoning models (LRMs) – an advanced form of AI – faced a “complete accuracy collapse” when presented with highly complex problems.

It found that standard AI models outperformed LRMs in low-complexity tasks, while both types of model suffered “complete collapse” with high-complexity tasks. Large reasoning models attempt to solve complex queries by generating detailed thinking processes that break down the problem into smaller steps.

The study, which tested the models’ ability to solve puzzles, added that as LRMs neared performance collapse they began “reducing their reasoning effort”. The Apple researchers said they found this “particularly concerning”.

Gary Marcus, a US academic who has become a prominent voice of caution on the capabilities of AI models, described the Apple paper as “pretty devastating”.

Writing in his newsletter on Substack, Marcus added that the findings raised questions about the race to artificial general intelligence (AGI), a theoretical stage of AI at which a system is able to match a human at carrying out any intellectual task.

Referring to the large language models [LLMs] that underpin tools such as ChatGPT, Marcus wrote: “Anybody who thinks LLMs are a direct route to the sort [of] AGI that could fundamentally transform society for the good is kidding themselves.”

The paper also found that reasoning models wasted computing power by finding the right solution for simpler problems early in their “thinking”. However, as problems became slightly more complex, models first explored incorrect solutions and arrived at the correct ones later.

For higher-complexity problems, however, the models would enter “collapse”, failing to generate any correct solutions. In one case, even when provided with an algorithm that would solve the problem, the models failed.

The paper said: “Upon approaching a critical threshold – which closely corresponds to their accuracy collapse point – models counterintuitively begin to reduce their reasoning effort despite increasing problem difficulty.”

The Apple experts said this indicated a “fundamental scaling limitation in the thinking capabilities of current reasoning models”.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

The paper set the LRMs puzzle challenges, such as solving the Tower of Hanoi and River Crossing puzzles. The researchers acknowledged that the focus on puzzles represented a limitation in the work.

The paper concluded that the current approach to AI may have reached limitations. It tested models including OpenAI’s o3, Google’s Gemini Thinking, Anthropic’s Claude 3.7 Sonnet-Thinking and DeepSeek-R1. Anthropic, Google and DeepSeek have been contacted for comment. OpenAI, the company behind ChatGPT, declined to comment.

Referring to “generalisable reasoning” – or an AI model’s ability to apply a narrow conclusion more broadly – the paper said: “These insights challenge prevailing assumptions about LRM capabilities and suggest that current approaches may be encountering fundamental barriers to generalisable reasoning.”

Andrew Rogoyski, of the Institute for People-Centred AI at the University of Surrey, said the Apple paper signalled the industry was “still feeling its way” on AGI and that the industry could have reached a “cul-de-sac” in its current approach.

“The finding that large reason models lose the plot on complex problems, while performing well on medium- and low-complexity problems implies that we’re in a potential cul-de-sac in current approaches,” he said.

‘Pretty devastating’ Apple paper raises doubts about race to reach stage of AI at which it matches human intelligence

Advanced AI suffers ‘complete accuracy collapse’ in face of complex problems, study finds

Apple’s artificial intelligence features took a backseat on Monday at its latest annual Worldwide Developers Conference. The company announced a revamped software design called Liquid Glass, new phone and camera apps as well as new features on Apple Watch and Vision Pro. But in spite of pressure to compete with firms that have gone all-in on AI, Apple’s AI announcements were limited to incremental features and upgrades.

Users will have a few new Apple Intelligence-powered features to look forward to including live translation, a real-time language translation feature that will be integrated into messages, FaceTime and the Phone app. The Android operating system has offered a similar feature for several years. Apple also introduced a new fitness app called Workout Buddy, which uses an AI-generated voice to speak to you during your workouts.

Consumers might soon notice some AI improvements to the non-Apple apps on their phone as well. The company said it was enabling app developers to tap into Apple’s on-device large language model to improve their experience with AI features within third-party apps. Consumers will be able to choose whether they want their data or information shared off-device and with the developers.

At last year’s WWDC, Apple announced a suite of AI upgrades to Siri that were intended to make the virtual assistant more personable and dynamic. Many of those features have yet to be released in spite of specific commitments from Apple.

“This work needed more time to reach our high-quality bar,” Craig Federighi, Apple’s vice president of software engineering, previously said of the delay.

The silence on Siri was “deafening”, wrote Forrester’s VP principal analyst Dipanjan Chatterjee.

“The topic was swiftly brushed aside to some indeterminate time next year. Apple continues to tweak its Apple Intelligence features, but no amount of text corrections or cute emojis can fill the yawning void of an intuitive, interactive AI experience,” Chatterjee wrote. “The end of the Siri runway is coming up fast, and Apple needs to lift off.”

Apple also announced a partnership with ChatGPT, a play to help the iPhone-maker catch up OpenAI, Microsoft and Google in the AI race. Wedbush Securities analyst Dan Ives said he expects Apple may need to forge more relationships with outside players to catch up with its competitors.

“Overall, WWDC laid out the vision for developers but was void of any major Apple Intelligence progress as Cupertino is playing it safe and close to the vest after the missteps last year,” Ives said.

skip past newsletter promotion Sign up to TechScape Free weekly newsletter A weekly dive in to how technology is shaping our lives Enter your email address Sign up Privacy Notice: Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Newsletters may contain info about charities, online ads, and content funded by outside parties. For more information see our Privacy Policy . We use Google reCaptcha to protect our website and the Google Privacy Policy and Terms of Service apply. after newsletter promotion

“We get the strategy but this is a big year ahead for Apple to monetize on the AI front, as ultimately Cook and co may be forced into doing some bigger AI acquisitions to jumpstart this AI strategy.”

AI announcements at WWDC limited to incremental features and upgrades despite pressure to compete