BBC News technology

Tech Now

Sat, 16 May 2026 01:00:00 GMT

X pledges quicker action on hate and terror content in the UK

Fri, 15 May 2026 12:57:30 GMT

Reuters

Laura Cress

Technology reporter

15 May 2026

Social media platform X has pledged to review UK reports of suspected illegal hate and terrorist content within 24 hours on average, under commitments accepted by Ofcom.

The Elon Musk-owned company said this would apply to content flagged through its illegal content reporting tool.

Ofcom's online safety director Oliver Griffiths called the commitments a "step forward" and said they were of particular importance after recent religiously-motivated crimes targeting Jewish communities in the UK.

Ofcom said a separate investigation into X's AI tool Grok, over concerns it was used to create sexualised images, is ongoing.

The BBC has contacted X for comment.

The announcement follows the launch of an Ofcom compliance programme in December, assessing whether the biggest social media companies have adequate systems and processes for dealing with reports of illegal hate and terror material.

Griffiths said the regulator had evidence that terrorist content and illegal hate speech was "persisting on some of the largest social media sites".

He said Ofcom was challenging the platforms to tackle the issue and take firmer action.

X will submit performance data to Ofcom every three months for a year, so the regulator can monitor whether it is meeting the targets.

While it has set a target to average less than 24 hours for its reviews, it has also promised to assess at least 85% of reports within 48 hours.

Ofcom also set out two further commitments from X to better protect UK users from illegal hate and terror content.

Under the first, the company will engage with experts about reporting systems for such content.

Ofcom said this followed concerns from some organisations that they had flagged "multiple pieces" of suspected illegal hate and terrorist content to X, but were unclear whether the reports had been received or acted upon.

X's second commitment is to withhold UK access to accounts reported for posting UK illegal terrorist content, if it determines they are operated by, or on behalf of, a terrorist organisation proscribed in the UK.

Getty Images

King Charles III met several members of the Jewish community in Golders Green following an attack in the neighbourhood

Danny Stone, chief executive of the Antisemitism Policy Trust, said the action was a "good start" but that there was still more to do.

"X is failing in so many regards to tackle open racism on its platform," he said.

"We know where this online harm leads, and so for the sake and safety of all of us in Britain, I hope Ofcom will hold X to account for what it has promised the regulator it will do."

The UK has seen a series of recent attacks targeting Jewish communities, including the Heaton Park Synagogue attack in Manchester in October 2025, an attack in Golders Green in April, and recent arson attempts on Jewish sites in London.

Iman Atta, director of Tell Mama, a national project which records anti-Muslim incidents in the UK, welcomed the updated targets, saying they signalled "a more accountable approach".

She said the group was "particularly encouraged" by the commitment to take action against accounts operated by or on behalf of terrorist organisations proscribed in the UK.

"This sends an important message that no platform or body operating in this country is above scrutiny," she said, adding that the test was "not what is promised, but what is delivered".

Police probe Islamist group's arson attack claims
X to stop Grok AI from undressing images of real people after backlash
Third arrest over arson attack at former synagogue

Claim, counter-claim and tech's seedy side exposed: Five things we learned in the Musk-Altman trial

Fri, 15 May 2026 05:03:31 GMT

To play this video you need to enable JavaScript in your browser.

This video can not be played

Watch: BBC explains closing arguments of the Musk-Altman trial

Lily Jamali

North America technology correspondentOakland, California

15 May 2026

It is the legal showdown that has pitted two of the biggest names in tech, Elon Musk and Sam Altman, against each other.

At stake is the future of one of the world's most valuable start-ups, ChatGPT-maker OpenAI, along with the reputations of Altman - the company's boss - and Musk, the man he founded it with.

The central claim the jury has now retired to consider is Musk's argument his former friend "stole a charity", cheating him out of a fortune (albeit a tiny one, by Musk's standards) along the way - something Altman strongly rejects.

But there's been much more to the trial than that.

Over the past three weeks, other reporters and I have been glued to our seats at the federal court in California as the evidence ranged from explosive text messages to revelations of free Teslas allegedly offered in exchange for power.

It has all been presided over by a no-nonsense judge who will take the jury's decision under advisement, but ultimately decides which side prevails.

For those who have been unable to follow every twist and turn, here are five big things we learned from the court battle.

1. It was more than Altman v Musk

Elon Musk's central claim in this lawsuit is that Altman lied to him about his commitment to OpenAI's non-profit status.

But this trial has ended up being more than one very famous man's word against another's.

A parade of witnesses - many of them also very high-profile figures in the world of tech - took to the stand during the trial and said they had never heard of or seen evidence of any such commitment from Musk himself.

Witnesses included OpenAI co-founder Ilya Sutskever, former OpenAI board member, Tasha McCauley, and even Microsoft boss Satya Nadella - who insisted his company did extensive due diligence before investing billions into OpenAI.

Of course, Nadella has skin in the game. Microsoft is a co-defendant, accused by Musk of aiding and abetting in Altman's alleged scheme.

Nevertheless, it was striking that this trial has not just been about Musk against Altman - there has been a barrage of voices contradicting the claims of the world's richest man.

'I thought he was going to hit me,' OpenAI co-founder says of Musk

2. Altman's character came under scrutiny

Reuters

Altman's turn on the stand, sketched by artist Vicki Behringer

Being able to call on some high-powered support from the witness stand hasn't stopped the questions about Altman's trusthworthiness.

In the weeks leading up to the trial, Altman had been the subject of a blistering New Yorker magazine profile by investigative reporter Ronan Farrow.

Zoning in on his career and moments such as his dramatic ousting from OpenAI in 2023, the story portrayed Altman as a pathological liar.

Musk's lawyer Steven Molo made the most of this.

"Are you completely trustworthy?" he asked Altman in his first question of cross examination.

Altman responded with: "I believe so." Molo ran with it.

"You don't know whether you're completely trustworthy?" he queried.

While the OpenAI boss asked to amend his answer to "yes", his character remained under a harsh spotlight throughout the trial.

Jurors heard from former OpenAI board members and executives - some in person, others via videotaped sworn depositions - detailing first-hand experiences of Altman allegedly failing to be forthright.

They also learned of his extensive investments in private start-ups, some of which have brokered deals with OpenAI.

A power purchasing agreement with nuclear energy start-up Helion Energy was flagged as particularly worrisome, given the firm has yet to ever deliver any power.

Altman was, until recently, chairman of Helion's board and holds a stake worth more than $1.5bn.

Musk accuses Altman of stealing a charity as courtroom battle begins

3. The no-nonsense judge didn't allow time for lunch

Musk v Altman wasn't just about the tech bros.

Other colourful characters who got their star turn turning this trial included Judge Gonzalez Rogers, who oversaw it.

She kept jurors, lawyers and media on a strict schedule, with only two twenty-minute breaks per day and no lunch time to ensure maximum alertness.

This judge had complete command of her courtroom and had no qualms about lambasting anyone who ran afoul of the rules.

That applied to observers who deigned to take photographs of the case's famous players in the courthouse and lawyers who pushed their questions into territory the judge had previously made clear were off-limits.

But she wasn't all business.

When the court experienced audio issues early in the trial, she addressed it and then deadpanned: "What can I tell you? We are funded by the federal government."

The drama of the trial, which could not be streamed in video, was also brought to life throughout by sketch artist Vicki Behringer, who painstakingly captured it in water-coloured splendour each day.

The no-nonsense judge calling the shots in Musk v Altman trial

4. Things at OpenAI got very personal

Getty Images

Once upon a time, Musk was Altman's hero.

The deterioration of that relationship was not the only highly personal situation the court focussed on.

Musk was generally confident, even combative, in his appearance as the first witness to take the stand.

But he got noticeably flustered when asked about his relationship with Shivon Zilis.

"We live together and she's the mother of four of my children," he acknowledged.

He described Zilis, an executive at his firm Neuralink, as one of his senior advisors.

She told the court that while also member of the OpenAI board, the billionaire had offered her his sperm after noticing she had no children - not your everyday boardroom interaction.

The fact Musk was to be the father to her children was something she, later, did not reveal to OpenAI colleagues until a media report was imminent.

An intriguing witness, she was almost robotic-sounding in her answers on the stand.

Her demeanour seemed warmer over text, where her job as 'Elon whisperer' came across most clearly.

She left the board after Musk started xAI.

"When the father of your babies starts a competitive effort and will recruit out of OpenAI, there is nothing to be done," Zilis wrote to a friend.

5. Free Teslas and frantic texts: how power really works in big tech

Getty Images

Protestors seized on the opportunity to blast Musk and Altman outside the Oakland courthouse

For outsiders, Musk v Altman offered a crash course in how power is wielded in Silicon Valley.

Want to lowball your co-founders? Give them free Teslas! (That was Elon, allegedly.)

Looking to ensure loyalty? Pay your most important strategic partner on the side. (That was Sam, allegedly.)

Musk's attorneys tried to diminish Altman before the jury, painting him as someone who tried to use a connection to Musk to bolster his own status.

In turn, Altman claimed Musk suggested control of OpenAI should go to his children.

Text messages divulged during the trial, external also put power struggles behind closed doors on full display.

These ranged from revealing Altman's frenzied response to his abrupt sacking in 2023, asking a former colleague at one point - "still don't want me?"

That same colleague in turn described how Altman was being replaced by Twitch boss Emmett Shear, or as she called him in her message "rando Twitch guy".

The flippant-sounding messages - as well as seeing these larger-than-life personalities sipping lattes around the courthouse - could make them seem a little less important.

Yet they still control technology that impacts the lives of billions of people. And they are embroiled in a row worth billions of dollars.

It is now over to the jury - and ultimately Judge Gonzalez Rogers - to decide what happens next.

Musk accuses OpenAI lawyer of trying to 'trick' him in combative testimony
What is AI and how does it work?

Inside the secretive and lucrative world of orchid breeding

Thu, 14 May 2026 23:03:16 GMT

Floricultura

The orchid industry is worth hundreds of millions of dollars

Matthew Kenyon

Technology ReporterHeemskerk, Netherlands

15 May 2026

It can take a decade of hard work to bring a new orchid to market.

While the rewards can be significant - the global orchid market is worth hundreds of millions of dollars - the competition to produce the next gorgeous flower is intense.

Which is why, in the race to develop new orchid types, the laboratory is at least as important as the greenhouse.

Centuries of human intervention - selective breeding and propagation - have made the genetic background of many commercial orchids a "disaster", according to leading Dutch orchid breeding firm Floricultura.

That means it is extremely difficult to predict what characteristics a new plant breed might have.

But by developing genetic markers for particular traits - colour, shape, disease resistance, flowering longevity and so on - Floricultura and its competitors can try to speed up the process of selective breeding.

Instead of waiting for a newly bred plant to flower in three years' time, the breeders can apply genetic screening techniques on very young plants and discard the ones that don't match their requirements, right at the start of the process.

"If a few thousand cross breeds [come] from the lab, we can screen them based on the marker and just select the ones that have the marker that you search for," says Wart van Zonneveld, Floricultura's research and development manager.

"It's an indication of a certain trait that you want or you do not want, depending on what's easier to find."

So-called "novel breeding techniques" are a closely-guarded secret. Each company develops its own genetic markers and processes because that's what allows them to develop unique varieties.

"We keep it to ourselves because it's lots of investment," van Zonneveld says.

"It's still breeding, you have to make a cross, and we cannot just pick out a piece of DNA and put it back that easily," says Paul Arens, ornamental plant breeding researcher at the Netherlands' Wageningen University.

He and his colleagues have carried out research for a Dutch government backed initiative that shares information with participating companies.

"The foundation is still what we are doing for 100 years already. You take two plants, you look at their characteristics, and you make a cross. But [the breeders] have white lab coats, [and] they're doing all kinds of research with markers, with genomics, on plant health."

Floricultura

Genetic markers are used to identify favourable traits in orchid plants

Genetics are also used in protecting the intellectual property in the new variety itself – in Europe through breeders' rights, and patents in the United States.

"If a company makes a new orchid, then [it] would like the sole right to commercialize this orchid," Arens says.

"Otherwise, somebody else can just buy it in the shop, multiply it and sell it himself.

"But the breeders' rights researcher has to make sure that a new variety is distinct from anything that's already in the market... it has to be distinct, it has to be stable, and it has to be uniform."

Breeders' rights and patents are granted based on physical descriptions, not DNA analysis, but it's essential to compare the new plants with similar products in order to establish if they qualify for protection.

DNA analysis is a powerful tool in determining which plants the new variety should be compared with.

"It's just like what we do in forensic science. You run markers that are at different positions in the DNA and that gives you a pattern and then you have a chance to match it or not," Arens says.

There is still an element of gambling to breeding orchids says Stefan Kuiper

Floricultura don't sell to the public, or even to garden centres. Their business is to produce and develop new varieties which they sell to the cultivators who grow the plants at scale.

They have more than 180 varieties in their catalogue, but several hundred more in development, because the demand for novelty and development never ceases.

"You can't stop, because it takes so long to develop new varieties," says Stefan Kuiper, the company's breeding manager.

"You have to go on, [or] you will be behind the rest."

After genetic screening and initial selection, the plants (the first attempts at a new variety, siblings from the parent orchids) take around three years to grow, first in lab conditions and then in greenhouses, but there are still years to run in the development stage.

Breeding, says Paul Arens of Wageningen University & Research, "is the art of throwing away", discarding those plants which don't match your ambitions, but it also the art of multiplying what's left.

Because the next batch of plants won't be siblings; they will instead be exact copies of the ones which survived the selection round - clones.

"In the beginning, everybody had the seedlings, so the crossing and then the seed pods give plants, but we at Floricultura introduced meristems," Stefan Kuiper tells me.

Meristems are the cells which allow a plant to continue growing through its life, and it's these which are used to clone the surviving plants.

Stefan can't explain more about the technique they use – like the genetic research it's a trade secret.

However, the cloned seedlings are cultivated and grown, again over years, to another selection point.

Floricultura

Young plants are sent to Poland and India for evaluation

The cultivation of orchids is a resource heavy business. The plants need reliable heat, light, water and nutrients, over many months.

The application of genetics and other techniques can only speed it up so far. Ultimately you have to let the plant grow, confirm the characteristics – flower shape and size, colour, the number of stems, resistance to disease and so on – and then make another selection.

That process takes the young plants by airfreight to India, and by lorry to Poland, before they return to the Floricultura site in Heemskerk in North Holland, where there are more than seven hectares of greenhouse space for both development and production.

Rainwater is harvested from the greenhouse roofs and in response to changing weather patterns, the company are now beginning to recycle that water, and the nutrients it contains, for secondary use.

Wart van Zonneveld proudly showed me their geothermal well, which pumps water up from 3km below the surface, at a temperature of 102C.

It provides so much energy that they are exploring sharing it with the local council for district heating projects.

It's not only the monitoring that is automated. In the vast greenhouses, trays of plants shuttle around on rollers, which deliver them to the next stage of cultivation in sequence.

There remains one task which, at Floricultura at least, is reserved for humans.

Whilst the tools for developing new varieties, cloning new plants and assessing the results have all been transformed by technological innovation, the decision on which varieties, after nine years of work, make it into the catalogue is still made by Stefan Kuiper and his colleagues in person.

A plant can have tick all the genetic boxes and produce all the right traits, but it has to be beautiful to sell – and that's a judgement made by people.

"Breeding is a little bit [like] gambling", Kuiper says, and for now that human element remains.

More Technology of Business

The 'dumb machine' promising a clean energy breakthrough
Published
21 April
Quantum computing: A tech race Europe could win?
Published
14 April
Could a digital twin make you into a 'superworker'?
Published
17 April

AI could put people off tech jobs and hurt the economy, warns Raspberry Pi boss

Thu, 14 May 2026 23:01:20 GMT

Getty Images

Alex Daniel

Business reporter

15 May 2026

The founder of British computer maker Raspberry Pi has warned that overestimating the abilities of Artificial Intelligence (AI) could put people off pursuing tech jobs and hurt the economy.

Eben Upton told the BBC's Big Boss Interview podcast this could "distort people's choices in ways that make that skill shortage worse and not better".

Some people are "very inclined to overestimate what these [AI] tools can do," he said, and warned against claims that it would destroy vast numbers of computing roles over the coming years.

The rise of tools such as ChatGPT and Claude have led to predictions of huge job losses, particularly for tech workers and graduates. Amazon, Meta and Microsoft have already blamed tens of thousands of layoffs on AI over the last year.

But some experts have suggested the technology is being used as a scapegoat for reducing headcount after a post-Covid hiring spree by many big corporates.

Raspberry Pi

Overestimating chatbots' ability to replace people could "undo a lot of the good work that's been done, not just by Raspberry Pi, but by a lot of other organisations" in encouraging people into tech careers, Upton said.

He added: "It's possible to get caught up in this. This is the risk of damage right in this moment of incredible enthusiasm for what are genuinely incredible tools."

"You read in the paper: 'What guidance should you give your child about what GCSEs to choose in the context of an AI future?' We have no data to inform a rational decision on that.

"The answer is: wait five years, wait 10 years, and then maybe we might know something."

Asked if the phenomenon could damage economic growth, he added: "Absolutely. We need a supply of engineers."

Big Boss Interview: Raspberry Pi's Eben Upton

Eben Upton, founder and CEO of Raspberry Pi, tells Felicity Hannah he is warning against claims that AI would destroy vast numbers of computing roles over the coming years.

Listen on Sounds

Raspberry Pi's devices are the most widely sold computers by a UK firm and are popular with people who do programming as a hobby.

Upton founded the company in 2012 because he was concerned young people were no longer getting computing skills because mobile phones and games consoles were replacing devices that they easily programme.

Raspberry Pi floated on the London Stock Exchange in 2024 and has become a success story for the UK stock market amid a swathe of other firms snubbing it for a listing in the US, including Cambridge-based chip maker Arm.

Upton said that while the UK had "enormous" industrial capacity, he admitted that high energy costs are "a challenge" for firms.

Britain has had among highest energy costs of the G7 nations in recent years, which has proved damaging for businesses.

"About the only reason I wouldn't do engineering build objects in the UK is the high cost of energy, and we need to do something about that," Upton said. "We're extremely lucky we're not running a fertiliser factory or an oil refinery."

"The cost of energy in the home has an impact on your labour cost. You have to pay people enough money to live."

'Millions' of pounds saved by replacing Palantir tech in refugee system

Thu, 14 May 2026 21:09:02 GMT

Getty Images

Chris Vallance

Senior technology reporter

14 May 2026

Millions of pounds have been saved by replacing a Palantir IT system which helps to find homes for Ukrainian refugees with one built by its own experts, a government department has said, external.

The Homes for Ukraine scheme matched people fleeing the conflict with offers of accommodation - a complex task Palantir initially supported for free but which grew to cost millions.

The Ministry of Housing, Communities and Local Government (MHCLG) said its new system was "more flexible" and could meet "high standards" of security.

Palantir said it was proud to have supported the scheme and "stood up a solution in just nine days, which enabled the safe resettlement of more than 157,000 refugees".

Homes for Ukraine was set up in in March 2022, less than a month after Russia's full scale invasion.

Through a website, backed by an IT system, those who had a rent-free space in their home or a separate residence could to offer it to refugees.

In order to set this up quickly, then-Conservative government ministers accepted an offer from Palantir to build a system to administrate the scheme, based on its Foundry platform, for free for six months.

In a 2023 blog post, external, Palantir described the challenge of combining data from multiple government systems containing tens of thousands of visa applications and hundreds of thousands of accommodation offers.

Subsequent 12-month contracts were awarded - one worth £4.5m and another £5.5m, according to a National Audit Office report, external.

The report notes the Government's chief commercial officer informed Palantir of his concern about the firm's practice of offering a zero- or nominal-cost initial offer to gain a commercial foothold.

This, he argued, was contrary to public procurement principles requiring open competition.

Palantir maintains government guidance suggests running pilots of systems and asking if they can be supplied for free.

The NAO report also said there was a desire to replace the Palantir system.

Dnipropetrovsk Regional State Administration

The current phase of the Russian invasion continues to destroy buildings and kill people in Ukraine

Coco Chan, a senior digital leader of the Homes for Ukraine project, said in a blog a system built on an existing commercial platform had been replaced with one created in-house.

The blog did not name the platform in question, now known to have been Palantir's Foundry tech.

"Longer term, we wanted to replace the platform with a more flexible technology solution, enabling [MHCLG] to save significant support costs, control the system data and code," Chan wrote.

She added its in-house replacement was "already saving MHCLG millions of pounds a year in running costs".

Towards 'sovereign technology'

According to Chan, the department set a precedent by moving a complex live system to an in-house set up, reducing reliance on external suppliers.

That message may be particularly welcome to those who have criticised Palantir and its contracts across UK public services - including with the NHS, the Ministry of Defence (MoD), the Financial Conduct Authority and 11 police forces.

Some argue the firm's success is because its tech is badly needed and works well.

But others contend Palantir's involvement with US immigration enforcement and Israel's military, as well as the beliefs of its two most prominent founders, make it an unsuitable partner.

There are also concerns the UK is relying too much on large US tech suppliers.

Terence Eden - who alerted the BBC to the MHCLG blog - said the development of an in-house alternative to Palantir's tech was an important step towards more "sovereign technology".

"When given suitable resources the Civil Service can often outperform private companies like Palantir," the former government technology advisor said.

Eden added MHCLG had created a "better, easier to use, and cheaper" system.

Palantir defends its record as MPs demand more scrutiny of data use
Palantir UK boss says it's up to militaries to decide how AI targeting is used in war

Emma Logan, deputy president of BCS, The Chartered Institute for IT, told the BBC there were clear advantages to building some digital services in-house.

But she said "external specialists can bring experience, specialist skills, and the ability to put large teams in place quickly, which can be particularly important for urgent national programmes".

Rob Miller, of Public Digital - a consultancy founded by former government tech experts - added that the government should not just consider whether to curb reliance on big tech "but how quickly it is willing to invest in the capability to do so."

Palantir told the BBC its Homes for Ukraine system formed part of a "multi-faceted effort to help Ukraine in the face of Russian aggression".

It added this included "the use of our software for military support, demining, investigation into war crimes and provision of pupils with safe access to schools".

The firm also said the change to a new system showed there was no risk of firms being locked into using its own.

The MHCLG said it initially needed a system which could be ready within days but, in seeking a "steadier service", later created an updated platform to meet the programme's longer-term needs and bring down costs.

Its replacement system was operational by September 2025.

HMRC to use AI from British tech firm to spot fraud and tax return errors

Thu, 14 May 2026 05:10:45 GMT

Getty Images

Zoe Kleinman

Technology editor

14 May 2026

HM Revenue and Customs has announced a 10-year, £175m deal with the British tech firm Quantexa to provide AI-powered technology to help improve its performance.

Quantexa says its systems will combine data collected by HMRC with external sources to help the tax office identify incidents of fraud and fix unintentional errors more quickly.

Its tasks will include helping HMRC to assist customer service staff, as well as to identify hidden networks of companies and individuals masking fraudulent activity.

Public dissatisfaction with HMRC performance has crept up in recent years, according to government figures.

A Freedom of Information request made by the campaigners at the Contentious Tax Group found there were more than 93,000 complaints made about the department in 2024-2025, external.

This was an increase from just over 70,000 in 2020-21, with one of the main gripes about poor response times.

Quantexa says automated decisions about tax payers made by AI will still need to be checked by people.

Quantexa chief executive Vishal Marria told the BBC the new technology was designed to "support human decision-making, not replace it".

"In government environments, AI cannot operate as a black box. Decisions need to be transparent, auditable, and explainable, particularly in areas affecting citizens directly," he said.

He also said HMRC data would remain secure, and staff working with the government department would remain separate from the rest of the business.

"We never take HMRC data away from the HMRC environment," he added.

Bloomberg via Getty Images

Vishal Marria is founder and chief executive of Quantexa

The company also said it will help in tracking down legitimate payments made to HMRC made under the wrong reference number.

The UK-based firm has been valued at $2.6bn (£1.9bn) and its corporate customers include HSBC and Vodafone.

The appointment of a British company is in line with the government's efforts to build what's known as "digital sovereignty".

There are concerns about the UK's dependence on platforms and services provided by big tech companies based in the US, including the controversial £330m contract awarded to the AI data processing firm Palantir to build a platform for the NHS.

AI minister Kendall says she doesn't use AI at work
OpenAI and UK sign deal to use AI in public services
The viral manifesto of 'anti-woke' tech boss with NHS and defence contracts

Thousands of Waymos recalled after robotaxi swept into a creek

Wed, 13 May 2026 14:40:20 GMT

Gado via Getty Images

Laura Cress

Technology reporter

13 May 2026

Waymo is recalling thousands of its self-driving cars in the US over a software issue that could allow vehicles to drive into flooded roads.

According to a letter posted on the National Highway Traffic Safety Administration (NHTSA), external website on Tuesday, the voluntary recall affects nearly 3,800 robotaxis that use the company's fifth and sixth-generation automated driving systems.

It follows an incident on 20 April in San Antonio, Texas, where an empty Waymo vehicle entered a flooded road and was swept into a creek.

The company, which hopes to operate a robotaxi service in London by September, told the BBC safety was its "primary priority" and it was working on "additional software safeguards".

A spokesperson from Waymo, which is owned by Google's parent company Alphabet, added "mitigations" had already been put in place, such as "limiting access to areas where flash flooding might occur".

Waymo's San Antonio service also remains temporarily suspended following the incident, though the company said it will resume public rides after the necessary software fix had been rolled out.

Waymo says it now provides more than 500,000 trips per week, external across multiple US cities including San Francisco, Austin and Miami.

Jack Stilgoe, professor of science and technology policy at University College London, told the BBC that all self-driving car systems had limits on when and where they could operate safely.

"We often see these limits only when something goes wrong," he said.

Getty Images

Waymo says it now provides more than 500,000 trips per week across multiple US cities

As more autonomous vehicles are deployed, Prof Stilgoe said, more such problems are likely to emerge.

"That isn't to say the technology won't be hugely beneficial," he added.

"But policymakers would prefer to know about these things in advance rather than discovering them in hindsight."

Over the past year several incidents with different driverless car firms have raised concerns over robotaxi safety.

In December 2025, a large power outage in San Francisco led Waymo taxis to stop working around the city, causing significant disruption.

And in April, a mass Apollo Go robotaxi outage in the Chinese city of Wuhan caused at least a hundred self-driving cars to stop mid-traffic.

Driverless taxi firm sorry for early-morning noise
AI teachers and cybernetics - what could the world look like in 2050?
Tesla investigated over self-driving cars on wrong side of road

WhatsApp launches totally private 'incognito' conversations with its AI chatbot

Wed, 13 May 2026 14:00:58 GMT

Reuters

Zoe Kleinman

Technology editor

13 May 2026

WhatsApp has introduced private chats with its AI chatbot which not even the tech company will be able to read in a new "incognito" mode.

It means neither the user nor the AI's responses will be monitored if the feature is activated, and past conversations will disappear from the chat for the user.

Will Cathcart, the head of WhatsApp, said he felt people wanted to have private conversations with AI on sensitive subjects including health, relationships and finances and didn't want them to be accessible.

But a cyber security expert has told the BBC this could lead to a lack of accountability for WhatsApp if things go wrong, as they would have no access to chat history.

WhatsApp is owned by Meta, which also owns Instagram, Facebook and Messenger.

When Meta AI was added to WhatsApp last year, it was criticised by some users angry at not being able to turn it off.

But in May 2025, Meta boss Mark Zuckerberg said Meta AI had reached a billion users, external across its apps.

"We've heard from a lot of people that they feel some discomfort about sharing [personal] information with the company, yet they want the answers," Cathcart said of the latest announcement.

Currently most AI companies do store some data from chatbot use, and outside of businesses who pay premiums for enterprise accounts, that data can also be used to train future models of the product.

Zuckerberg described it as the "first major AI product where there is no log of your conversations stored on servers".

The technology behind WhatsApp's incognito mode is not the same as the end-to-end encryption the platform uses to protect other messages, but it is "the equivalent," Cathcart added.

Prof Alan Woodward, a cyber security expert at Surrey University, said there was a low risk of compromising WhatsApp's existing security by introducing a second system.

However there are concerns about how incognito mode could hide AI malfunction or abuse.

A number of AI companies, including OpenAI and Google, have been the subject of wrongful death lawsuits.

Woodward said there was a risk of a lack of accountability for the AI's responses.

"Personally I think what you ask an AI should remain private as some people ask it very personal matters - but you are placing a great deal of trust in the AI not to lead users astray," he told the BBC.

The concern is that disappearing messages which cannot be retrieved by the user or by Meta would mean it would be impossible to find evidence whether somebody's chats led to harm, death or suicide.

Cathcart said incognito mode would initially only process text rather than images, and Meta AI's guardrails would err on the side of caution in refusing to answer requests which could be interpreted as harmful or illegal.

WhatsApp has blocked other AI chatbots from being accessible from its systems, so the only AI that its billions of users can interact with on the platform is Meta's own.

"Meta is on track to shell out $145bn [£107bn] on AI infrastructure in 2026, and investors want to see a lot more bang for those mega bucks," said Susannah Streeter from investment platform Wealth Club.

"Investors are nervous about the scale of the spending, but Meta is counting on the push leading to significant returns. It's building out AI infrastructure to supercharge its platforms and if it works it could make its advertising and commerce empire even more dominant."

Meta shares slide as plan to spend billions more on AI spooks investors
Meta AI searches made public - but do all its users realise?
Meta investigated over AI having 'sensual' chats with children

Suicide forum fined £950,000 for not blocking UK users

Wed, 13 May 2026 11:52:58 GMT

PA Media

13 May 2026

Imran Rahman-Jones

Technology reporter

A pro-suicide online forum which has been linked to at least 50 deaths has been fined £950,000 by the UK's media regulator.

Ofcom said the site did not comply with the Online Safety Act (OSA) "to protect people in the UK from illegal content".

Ofcom Director of Enforcement Suzanne Cater said the forum had made some attempts to block UK users but this was "not good enough and the changes they've made were not consistently applied or effective to reduce the risk of harm".

But Ofcom has been criticised for taking too long to take action with the Molly Rose Foundation saying: "It is appalling that it has been left to bereaved families and campaign groups to press Ofcom into action."

The online safety charity's chief executive Andy Burrows added it had "submitted detailed evidence which showed scores of vulnerable young people remained at risk while Ofcom's investigation dragged on.

"Further lives were lost during this period".

Ofcom said it understood the urgency and the anger felt by those who had been personally affected.

"It is vital that we ensure our enforcement action is thorough, and this can take time, as is the case for any enforcement agency," a spokesperson told BBC News.

It is a criminal offence in the UK to intentionally encourage or assist suicide, and the OSA requires chatroom providers to mitigate the risks of UK users seeing this type of content on their platforms.

Ofcom said the site was accessible by people in the UK without the use of a VPN, which is a common way of getting around blocks in one country.

The regulator added it had spoken to people with "personal knowledge of the forum" and thanked them for engaging with the investigation.

Its investigation ran from March 2025 to April 2026 and was the first of its type under the OSA.

Ofcom said it had "engaged extensively" with the forum provider throughout.

To play this video you need to enable JavaScript in your browser.

This video can not be played

Vlad Nikolin-Caisley's family tearfully spoke to the BBC to call for tighter restrictions on online platforms

BBC News investigations into the forum found young people who had been on the site had been encouraged to and given instructions on how to kill themselves.

The BBC and Ofcom chose not to name the website.

Vlad Nikolin-Caisley, 17, and Aimee Walton, 21, both from Southampton, died after using the online pro-suicide chat room.

Their families had called on Ofcom to act.

After the fine was announced, Aimee's sister Adele Zeynep Walton said families like hers had been "agonisingly waiting for action against the website that took our loved ones".

In a statement on behalf of Families and Survivors to Prevent Online Suicide Harms, Walton added: "We feel let down by the process and Ofcom's slow response to this threat to life."

She also said a fine was not enough, and wanted to see "criminal sanctions against the sinister actors who actively groom, encourage and instruct British people to take their lives."

An Ofcom spokesperson said the provider had "failed - and continues to fail - to comply with its duties".

They added: "Our investigation into this forum was the first to be launched under the Online Safety Act straight after it came into force."

The forum had made some efforts to block UK users, but in November 2025 the BBC reported a "mirror" site had emerged under a new name.

Ofcom said it acted quickly after being alerted to the new site by the Samaritans charity, and it was blocked again.

Killed By The Internet

Suicide at seventeen. The tragic consequences of an online forum obsessed with dying.

Watch on iPlayer

Ofcom said the £950,000 fine "reflects the serious and deliberate nature of the contraventions, and the risk of fatal harm to people in the UK posed by the content present on the service".

It also took into account the changes which the forum's provider had made, as it had attempted some blocks to UK users.

However, it said the provider had failed "to comply with its duties to assess and mitigate the risk of people in the UK encountering illegal content on its service".

The provider now has 10 working days to comply with the law, while Ofcom is preparing to apply for a court order which would require internet service providers to block access to its site.

If the company fails to pay the fine, Ofcom can pursue it through the courts.

Cater said the forum had "caused unimaginable pain and suffering for bereaved families across the UK and beyond, and no punishment can undo that harm".

If you have been affected by any of the issues in this story you can find information and support on the BBC Actionline website here.

How can you keep your child safe online?
'Toxic online community led to my sister's suicide'
Suicide forum warned of potential £18m fine by Ofcom

More than 70 million warnings sent to people seeking child abuse material

Wed, 13 May 2026 06:04:14 GMT

Getty Images

Shiona McCallum

Senior Tech Reporter

13 May 2026

More than 70 million warning messages have been sent to people attempting to access child sexual abuse material (CSAM) online over the past two years, the Lucy Faithfull Foundation says.

The messages are sent as part of Project Intercept, external, a partnership between the child protection charity and technology firms including Google, TikTok and Meta.

Rather than simply blocking content, the messages highlight the illegality of viewing CSAM and direct users to support services aimed at changing behaviour.

The foundation said nearly 700,000 people went on to access its Stop It Now resources, which offer confidential advice and self-help tools - a figure some experts say is disappointingly low.

"Given that 70 million warning messages have been sent, the fact that only 700,000 people click through to get support seems low. This is disappointing, given that the scale of the problem of child sexual abuse imagery online is growing fast," said Professor Sonia Livingstone, director of the Digital Futures for Children centre at London School of Economics,

"On the other hand, since four in five of those people who seek support do engage with the resources provided, that suggests the system is working for those who are really motivated to get help."

Lucy Faithfull Foundation

Disrupt and divert

Project Intercept is active in 131 countries and operates across a range of online spaces.

These include end-to-end encrypted services - where only the sender and recipient can view what's sent - and AI chatbot platforms.

The foundation did not specify how many individual users were responsible for the searches.

But it said engagement with the support material had been high, with an average of 28,000 users a month redirected in 2024 and 2025.

More than four in five continued to interact with the content, although the organisation did not publish data on longer-term behaviour change.

Deborah Denis, chief executive of the Lucy Faithfull Foundation, said: "By placing warnings at the moment harmful behaviour is happening, we can disrupt it and divert people towards help to change," adding that the approach could be scaled even further.

Children's charity the NSPCC said that interventions of this kind could play an important role in disrupting harmful behaviour, but should form part of a wider set of measures aimed at stopping illegal material from being created and shared in the first place.

The child protection charity said tech companies needed to go further in tackling the spread of such content.

Emma Hardy, Communications Director at the Internet Watch Foundation, said "innovative solutions" needed to be looked at including on parts of the internet that are end-to-end encrypted.

"As it is, it is simply too easy to share and distribute child sexual abuse imagery online, and for children to become trapped in cycles of exploitation, she said.

"Safety by design needs to be a guiding principle and new products and platforms must be built to make sure there is nowhere for this sort of behaviour to hide."

Ofcom, the communications regulator, said warning messages formed part of its expectations under the UK's Online Safety Act.

Child Protection Policy Director Almudena Lara said the data highlighted both progress and "the scale of the problem that still needs to be addressed".

Tech firms involved said the approach complements existing moderation systems.

Griffin Hunt, a product manager at Google Search, said changes made in early 2025 had led to "greater engagement with therapeutic help services" and fewer follow-up searches for illegal material.

Mega - a company which sells encrypted cloud storage - is also involved in the project, which it said challenged the idea that encrypted services could not intervene early to address harmful behaviour.

How can you keep your child safe online?
Tech firms told to do better on child abuse images
Online porn showing choking to be made illegal, government says

Smart glasses are 'an invasion of privacy' - Meta's are selling better than ever

Wed, 13 May 2026 05:01:07 GMT

Reuters

Mark Zuckerberg, seen on stage talking about Meta's Ray-Ban smart glasses

Kali Hays

Technology reporter

13 May 2026

Issues with a new wave of "smart glasses" seem to be piling up.

Yet some of the biggest technology companies in the world are poised to sell many millions of pairs in the coming years.

Women leaving the beach, going into a shop, or simply standing outside are now being approached by men usually wearing Meta's Ray-Bans, the company's "smart" or "AI" glasses, often in order to film the women's responses to casual questions or pick-up lines without their knowledge or consent.

The women only find out about the videos of them after they gain traction, and often abuse, online. They have little legal recourse as photography in public is broadly considered legal. One woman told the BBC that when she asked the person who posted a secret recording of her to remove it, she was told that doing so was "a paid service".

Meta's glasses are currently the most popular on the market, estimated, external to make up more than 80% of all AI or smart glasses sales, as the company was the first major tech player to launch such a product in recent years.

Made in partnership with EssilorLuxottica and offering the classic look of Ray-Bans, the glasses feature an almost invisible camera in the frames, small speakers in the arms, and in the latest version lenses that can show a wearer some information. People can start recording video or take a photo with a casual touch of the frames.

The nature of the camera in Meta's glasses can be so unobtrusive that even their wearers have been caught off guard by what and when they're recording, and where those recordings are going.

After workers in Kenya, tasked with watching videos made through Meta's glasses to create AI training data for the company, said they were being required to watch graphic content like sex and bathroom usage, people who own the glasses filed two lawsuits. In one, people said they had no idea such videos had been made. In the other, they said they did not know their videos were being shared by the company for review.

Meta has previously said that users were made aware of the possibility of human review in some circumstances in its terms of service.

Nevertheless, sales continue to rise. Today, seven million pairs and counting have been sold, according to the company.

"They're some of the fastest-growing consumer electronics in history," Mark Zuckerberg, Meta's chief executive, boasted earlier this year.

Tracy Clayton, a Meta spokesman, told the BBC that people should behave responsibly with any technology.

"We have teams dedicated to limiting and combating misuse, but as with any technology, the onus is ultimately on individual people to not actively exploit it."

Now, other major tech companies are planning to get in on what may have the potential to be the tech industry's long-awaited new product category.

Apple is reportedly developing, external its own version of smart glasses, possibly to be released next year. Snap has said it will release a new version, external of its smart glasses, called Specs, this year.

Google, too, is set to try again, external with smart glasses, more than a decade after its notorious Google Glass flop, which the company pulled from the public within two years of launch as the pricey gadget came under fire over privacy concerns.

All are expected to offer some combination of artificial intelligence (AI) and augmented reality (AR) technology, as Meta's glasses do, which typically requires a camera.

Mark Smith

Mark Smith wearing his Meta Ray-Bans

The way people may use the coming wave of smart glasses will not be all bad, of course.

Mark Smith wears his Meta Ray-Bans every day.

"I've used them around the world, in all kinds of places. The basic features are great," Smith said.

As a partner at the advisory firm ISG where he focuses on enterprise software, Smith can be classified as a tech-savvy early adopter. But the reasons he likes the glasses are not about any huge leaps in technological capability.

He likes to wear them while washing up the dishes at home because they make it easy for him to listen to music or a podcast without blocking out other noise like most headphones do. Taking phone calls through the glasses is a breeze. When travelling, it's nice not to have to constantly pull out his phone to snap a quick picture or video.

Even so, Smith said some potential privacy issues are obvious. The small light that turns on when the glasses are recording appears dim in daylight and often goes unnoticed, he said. Most people seem to have no idea he's wearing anything other than normal eyeglasses.

Should AI or smart glasses products from more companies end up selling as well as Meta's version, researchers, external expect as many as 100 million people, external will buy a pair in the next few years.

If such a prediction becomes reality, the ability of institutions to enforce norms and laws that typically prohibit recording in places like courthouses, museums, movie theatres, hospitals and bathrooms will be difficult when suddenly millions of eyeglasses are also cameras.

David Kessler, an attorney who heads the US privacy practice at Norton Rose Fulbright, said many of his corporate clients are already having to grapple with this.

"There are some pretty dark places we could go here," Kessler said. "I'm not anti-technology in any sense, but as a societal matter...will I need to think [of being recorded] anytime I go out in public?"

And Meta reportedly plans, external to add facial recognition technology in an updated version of its glasses, meaning wearers could not only have the ability to surreptitiously record anyone, but quickly identify them, as well.

PA Wire

A Google executive wearing the Google Glass in 2013

Meta markets its glasses under the tagline: "Designed for privacy, controlled by you." It suggests to users of the glasses that they do not record people who state they do not want to be recorded, and that users turn the glasses off completely "in sensitive spaces".

Those suggestions often seem to go ignored.

An increasingly popular use of the glasses is to record pranks on unsuspecting people.

Wearers, often young men, get people to sign fake petitions or retail workers to smell candles sprayed with bad odours. Sometimes they steal food as it's being handed over at a drive-thru and record their sprint.

People do often recoil when they discover a person is wearing smart glasses.

Online influencer Aniessa Navarro said she felt sick when she realised during a personal waxing session that her technician was wearing Meta's glasses. The technician said they weren't charged or recording, and that she needed to wear them for the prescription lenses.

Andrew Bosworth, Meta's chief technology officer, was asked on Instagram two weeks ago about "the stigma around people wearing smart glasses every day".

He responded, external by saying that the sheer number of Meta Ray-Bans sold "suggest that these are widely accepted".

But David Harris, a former Meta AI researcher who now teaches at UC Berkeley and is an adviser on AI policy in the US and EU, said he expects this generation of AI smart glasses to run up against the same issues that doomed the Google Glass more than a decade ago.

"Technology like this is fundamentally an invasion of privacy and it's really going to face more and more backlash," he said.

More signs of such a backlash are starting to appear.

In December a man posted a video complaining that a woman he'd been recording on the New York City Subway broke his Meta glasses. If he was expecting sympathy he was wrong. The internet hailed her as a hero., external

Tech Life

Tue, 12 May 2026 19:45:00 GMT

Crumbling roads and other hazards - can AI make our streets safer?

Elon Musk said control of OpenAI should go to his children, Sam Altman tells jury

Tue, 12 May 2026 18:51:56 GMT

Reuters

OpenAI CEO Sam Altman

Lily JamaliNorth America Technology correspondent and Kali HaysTechnology reporter

12 May 2026

Elon Musk tried to take control of OpenAI, even suggesting it could pass to his children when he dies, Sam Altman said on Tuesday.

Altman is co-founder and chief executive of the artificial intelligence (AI) company behind ChatGPT. He is being sued by Musk, who accuses him of having "looted a charity" given OpenAI began as a non-profit.

Appearing before a federal jury in Oakland, California, Altman said Musk not only backed the idea of OpenAI becoming a for-profit business, he wanted control of it for the long-run.

"A particularly hair-raising moment was when my cofounders asked, 'If you have control, what happens when you die?' He said something like '...maybe it should pass to my children.'"

The alleged comments from Musk came as the billionaire was trying to get more control of OpenAI after it was founded in 2015, and had floated several ways of gaining it.

As well as wanting more seats on OpenAI's board and to become its chief executive, Altman said Musk suggested OpenAI become a subsidiary of Tesla, his electric vehicle company.

The overarching goal for OpenAI was getting "more money faster", Altman said on Tuesday, recalling conversations that involved Musk about the company restructuring to become a more traditional for-profit entity.

Musk allegedly felt he should be in charge of such a company, in part because of his fame as a business person who could help OpenAI get financial backing.

"If I make one tweet about this, it's instantly worth a ton," Altman recalled Musk saying.

But Altman said he, as well as OpenAI co-founders Greg Brockman and Ilya Sutskever, decided handing Musk such control in exchange for more or easier financing would not help OpenAI's mission or its pursuit of artificial general intelligence, or AGI.

"I was extremely uncomfortable with it," Altman said. "One of the reasons we started OpenAI was because we didn't think any one person should be in control of AGI."

AGI is loosely defined in the tech and AI community as an AI tool or model that becomes so capable and "intelligent" that it outperforms humans on most tasks.

Ultimately, Musk left OpenAI in early 2018 and stopped his quarterly donations of $5 million to the company.

Altman described as "burned into my memory" an email from Musk in which he said OpenAI "had a zero percent chance, not a one percent chance, of success" without him.

When Altman offered Musk the option of investing in OpenAI when it formed a for-profit subsidiary in 2019, Musk declined.

"He said no because he would no longer invest in any startups he didn't control," Altman said.

Canvas hack: Company pays criminals to delete students' stolen data

Tue, 12 May 2026 10:31:23 GMT

Getty Images

A stock image of US university students in exams

Joe Tidy

Cyber correspondent, BBC World Service

12 May 2026

The company behind the popular Canvas software, which was hacked last week causing major disruption at thousands of universities and colleges, has paid the hackers not to publish stolen data online.

The cyber-attack affected an estimated 9,000 institutions in the US, Canada, Australia and the UK, with exams disrupted after the Canvas service went down.

The hackers threatened to publish 3.5 terabytes of student and university data they had stolen in the breach.

Instructure, the maker of Canvas, has now confirmed it has "reached an agreement" , externalwith the hackers, who have said they deleted the data and promised not to extort any students or institutions.

Paying cyber criminals goes against the advice of law enforcement agencies around the world, as it can fuel further attacks and offers no guarantee the data has been deleted.

In previous cases, criminals have accepted ransom payments but lied about destroying stolen data, instead keeping it for resale.

For example, when the notorious LockBit ransomware group was hacked by the National Crime Agency, police found stolen data had not been deleted even after payments had been made.

Instructure said in a statement on its website that protecting students' and education staff data was its primary motivation.

"While there is never complete certainty when dealing with cyber criminals, we believe it was important to take every step within our control to give customers additional peace of mind, to the extent possible," the company said.

Instructure did not set out the terms of the agreement but said that it meant that:

the data was returned to the company
it received "digital confirmation of data destruction"
it had been informed that no Instructure customers would be extorted as a result of the incident
the agreement covers all affected customers, with no need for individuals to engage with the hackers

The breach was discovered on 29th April and was claimed online by the prolific Shiny Hunters extortion group.

Neither the hackers nor the company are explicitly saying that money was exchanged, but cyber extortion groups like Shiny Hunters operate by forcing their victims to send money in bitcoin after a negotiation through an encrypted chat service.

A ransom note demanding payment in bitcoin appeared on screens during a cyber-hacking incident on the cloud-based platform Canvas.

It is unusual for victims of cyber attacks to publicly acknowledge paying hackers, but Instructure has maintained a high level of transparency, providing regular updates on its website.

That openness may be partly because the attack was highly visible and affected students directly.

Students sitting exams in the US were particularly badly affected, losing access to Canvas for revision and, in some cases, having online exams interrupted.

Aubrey Palmer, a meteorology student at Mississippi State University, told the BBC that they and other students had just finished writing a 2,900‑word exam essay when a ransom message suddenly appeared on their screens.

Courtesy of Aubrey Palmer

Affected student Aubrey Palmer

The note read: "Shiny Hunters has breached Instructure (again)."

It threatened to release stolen data unless a ransom was paid in bitcoin by Canvas or affected universities.

"My knee‑jerk reaction was that I'd been hacked myself, because that's what it looked like," Aubrey said.

"But then I actually read the ransom note and saw it was Canvas that had been hacked."

Aubrey added dozens of students received the same message, and there was confusion in the exam room about whether their work had been saved.

Mississippi State University later announced some exams would be postponed to allow students to recover any lost work.

Shiny Hunters is known for hacking organisations, stealing data and then publicly pressuring victims to pay ransoms in bitcoin.

The group has been linked to other breaches, including attacks on Jaguar Land Rover and Gucci. The criminals are English‑speaking and believed to be young.

In Telegram messages exchanged with the BBC, Shiny Hunters said it had hacked Canvas twice before last Thursday's attack.

Instructure disclosed a breach in September 2025 in a post on its blog.

Shiny Hunters has also claimed it breached the company again in April 2026, ahead of the 29 April attack.

When asked how it felt about the stress and disruption caused to students like Aubrey Palmer, the group said: "We have no comment on that."

It would not say how much it had been paid by Instructure.

Have you been affected? Share your experiences here

Tech Now

Sat, 09 May 2026 01:00:00 GMT

Tech Life

Tue, 05 May 2026 19:30:00 GMT

A sleep apnoea wearable. Plus, tech to free up hospital beds and a breast health app.

Tech Now

Sat, 02 May 2026 01:00:00 GMT

Tech Life

Tue, 28 Apr 2026 19:32:00 GMT

Workers train AI to identify everyday objects by labelling items in photos and videos.

Tech Life

Tue, 21 Apr 2026 20:15:00 GMT

How to say goodbye and pay tribute to a loved one - with a hologram.

Tech Now

Sat, 04 Apr 2026 01:00:00 GMT

BBC News technology

Tech Now

X pledges quicker action on hate and terror content in the UK

Related topics

Claim, counter-claim and tech's seedy side exposed: Five things we learned in the Musk-Altman trial

To play this video you need to enable JavaScript in your browser.

1. It was more than Altman v Musk

2. Altman's character came under scrutiny

3. The no-nonsense judge didn't allow time for lunch

4. Things at OpenAI got very personal

5. Free Teslas and frantic texts: how power really works in big tech

Related topics

Inside the secretive and lucrative world of orchid breeding

Related topics

More Technology of Business

AI could put people off tech jobs and hurt the economy, warns Raspberry Pi boss

Related topics

'Millions' of pounds saved by replacing Palantir tech in refugee system

Towards 'sovereign technology'

Related topics

HMRC to use AI from British tech firm to spot fraud and tax return errors

Related topics

Thousands of Waymos recalled after robotaxi swept into a creek

Related topics

WhatsApp launches totally private 'incognito' conversations with its AI chatbot

Related topics

Suicide forum fined £950,000 for not blocking UK users

To play this video you need to enable JavaScript in your browser.

Related topics

More than 70 million warnings sent to people seeking child abuse material

Disrupt and divert

'Too easy to share'

Smart glasses are 'an invasion of privacy' - Meta's are selling better than ever

Related topics

More on this story

Tech Life

Elon Musk said control of OpenAI should go to his children, Sam Altman tells jury

Related topics

More on this story

Canvas hack: Company pays criminals to delete students' stolen data

Related topics

More on this story

Tech Now

Tech Life

Tech Now

Tech Life

Tech Life

Tech Now