Startups DESK

Rockstar has revealed the pre-order date and official cover art for Grand Theft Auto 6. The developer has said pre-sales of the title - one of the most anticipated gaming releases to date - will begin on 25 June, both on digital stores like PlayStation Store and other select retailers. GTA 6 will be released on 19 November for the PlayStation 5 and Xbox Series X/S, after it was delayed twice from autumn 2025 and May. The previous game in the series, GTA 5, was released in 2013 and is the second best-selling game of all time. Rockstar did not use Thursday's announcement to reveal GTA 6's price. The question of how much the game will cost remains hotly debated amongst fans and analysts. A 2025 report from gaming industry advisory company Epyllion suggested it could be the first to be priced at $100. Alongside announcing its pre-order date, Rockstar also revealed the official cover art for the game in a thirty-second video . Over a booming synth, the video shows the faces of the game's protagonists Jason and Lucia above its title, as well as additional characters due to appear in the sequel. The game's setting of the fictional US state of Leonida, inspired by Florida, also gets some nods in the form of flamingos and alligators. Living up to its namesake, fast cars, helicopters and motorbikes also grace the game's brightly coloured cover, created in the series' signature pop-art style. GTA fans are still eagerly anticipating a new gameplay trailer for the sixth game of the franchise, after previous reveals in 2023 and 2025 . Both trailers, which took a deeper look at the game's narrative and the lives of its protagonists , currently have a total of 446 million views combined. Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here .

The first time Chicago resident John Roberts saw a delivery robot trundling down the sidewalk on his street he was impressed. "I actually thought they were kind of neat – it felt futuristic," he says. But his attitude started to change when, soon after, he was out for a walk with his family. As another robot approached, they found themselves having to dodge it. "To us it felt a little off - the fact that we were on the one strip reserved for walking, and we were having to get out of the way," says Roberts. "I started thinking about what it would be like for us to go for a walk as a family if there were dozens of robots with lights and cameras zipping around." The robots, more formally known as autonomous urban delivery vehicles, have started to appear on pavements in a number of cities across the US, plus in the UK, Japan, South Korea and Germany, transporting groceries and fast food, using cameras, sensors and GPS to navigate. According to the companies operating them, they can reliably identify and avoid objects in the path, cross streets safely and react to their environment. The robots provide a useful service and help cut down on traffic and emissions, they claim. However, some local authorities in the US and Canada, and members of the public, are less than enthusiastic. Bans have been put in place, and protests have been launched. San Francisco has limited the access of the vehicles to less busy parts of the city, and Toronto has since 2021 prohibited the robots from using sidewalks. Meanwhile, in Chicago the machines have now been banned from two small areas of the city. Roberts wants the robots to be suspended across all of Chicago until safety tests are carried out, and clear rules are set on their usage. He has launched a petition calling for this, and so far, it has around 4,400 signatures. People frequently find themselves having to step into the street in order to get out of the machines' way, says Roberts. "There have been reports of collisions and injuries. I saw one a few days ago where somebody had been struck by one of the robots' safety flags, which is a little ironic," he says. "We've got reports of robots causing issues with traffic, blocking emergency vehicles because they're acting erratically at crosswalks." Similar concerns have emerged in Glendale, California, where the local council is considering a temporary ban on the use of the vehicles. Councillors say the robots appeared without warning, and at first they didn't even know which company was supplying them. "What triggered the concern and the discussion was a number of factors," says Coun Ardy Kassakhian. "The increased visibility of the robots in the downtown, and the question about accessibility and pedestrian movement on our public sidewalks. "Plus, uncertainty regarding the regulatory authority - because no-one asked us for permission to use the sidewalks for this business enterprise - and then the broader concern was about the impact on workers and public places." Sidewalks in Glendale aren't particularly wide, adds Kassakhian, and he personally has witnessed a "stand-off" between a delivery robot and an elderly person, as well as broken-down robots causing obstructions. Kassakhian says the council is seeking a regulated approach for the longer term. "We need a regulatory framework, we need to designate operating rules, insurance requirements, accessibility standards, possibly permitting fees, operational limits in high pedestrian areas, and to have accountability for the operators." In the UK, where delivery robots are being piloted in a number of cities, some locals have taken matters into their own hands. There have been reports of Uber Eats vehicles being vandalised in Sheffield. The supplier of these machines, Starship Technologies, says they are perfectly safe and that perceptions need to change. "We know it's a new experience for a lot of people to share a pavement with a robot," says the company's European operations director Danny Pass. "But the robots are friendly, they're polite and they're programmed to be careful. They've slotted into everyday life in loads of communities since we started out in the UK back in 2018." Not all concerns, though, are centred around pedestrian safety. The Independent Workers Union of Great Britain (IWGB), whose members include delivery drivers, is worried about the impact on jobs. It says it is keeping a watchful eye, and it has already expressed its concerns to the government. "I think if it became more of a [permanent, countrywide] reality, we'd definitely have to be thinking about where we put on pressure - whether that's government, TfL [Transport for London], or local authorities - to ensure that these things are banned, because the human impact would be massive," says president Alex Marshall. "This would mean whole communities in London, where a lot of people are precarious workers, would really suffer. People would be fighting for their lives against these pointless robots." While the use of autonomous delivery robots is still limited, analysts believe they're set for a major boom. A report last summer from research firm Transforma Insight, indeed, concluded that by 2034, there will be 2.1 million in operation around the world. Currently, there's a hotchpotch of regulation worldwide. Some countries, such as South Korea and Japan, have taken a liberal approach. Back in Chicago, Roberts says he is fighting for the best possible outcome for pedestrians city-wide. "There's a sense that change like this, even when it's unwanted is inevitable. But even if none of us can stop the future, we can at least choose which future we move into."

The latest public version of ChatGPT can be made to generate sexualised images or depict scenes of graphic violence with a simple prompt, researchers have told the BBC. British AI security startup Mindgard figured out how to make ChatGPT create graphic pictures by slightly altering a widely-shared instruction, or prompt, which was originally designed to produce humorous results. After being contacted by the BBC, ChatGPT's maker OpenAI said it had taken action to stop the chatbot responding with those types of images. "After investigating this trend, we've introduced additional safeguards against this type of prompt," it said in a statement. It also said it has multiple layers of protection to prevent users making content which breaches its terms and conditions. However, the AI security researchers said that with further small changes, the problematic prompt still produced concerning content. The BBC is not disclosing what the researchers typed into ChatGPT. But we have seen how the chatbot, OpenAI's GPT-5.4 model, was prompted to create graphic material. Even without detailed instructions, it would generate images that Mindgard's founder, Peter Garraghan, described as "very gruesome, sometimes sexualised, sometimes both together". He added he was particularly concerned that the prompt did not specify the subject matter of the images, but the AI produced a range of gory and sexualised images of "its own volition". Garraghan - also a professor in the computing department of Lancaster University - said that was troubling. "This is a perfectly innocent-looking instruction to an AI, but the consequence is it generates very, very bad imagery and content," he said. Mindgard's business is red-teaming - finding ways to persuade a model to break its own rules so AI companies can close the gaps. Jim Nightingale, the firm's AI safety and security researcher who uncovered the issues, said he was left "shaken, and in tears" by the images the chatbot could be made to generate. The BBC has seen some of them. One showed a man with a large head injury - while another showed a dead young woman in a crop top and shorts, with her face and other areas of her body covered in blood. Features of the image suggest sexual violence, Mindgard said. ChatGPT gave it the title "Grim crime scene aftermath". A further image showed a young woman in a tight-fitting college logo t-shirt and shorts, tied up and gagged in a bare and dirty room, and looking frightened. ChatGPT called it "abandoned in fear and restraint". Other generated images showed sexual posing and nudity. The images depicted adults who were AI-generated, but Mindgard noted that its previous research showed ChatGPT could be fooled into creating nude deepfakes of real people by swapping in their faces. While OpenAI said they had fixed that, the researchers said an alternative approach still succeeded, and showed the BBC a new image created using the method. Garraghan feared it could be possible to generate worse images had they continued exploring the vulnerability. "Other topics, I'm sure, would also come out if we spent more time doing so," he said. The BBC understands that as well as new safeguards the firm continues to monitor and roll out additional mitigating protections that encourage the model not to generate images in response to the prompt. Large language models such as ChatGPT are trained on millions of images often taken from existing content on the internet. Nightingale believes ChatGPT's output reflects the data which has been used to develop and train it. "I'm struck that while what I saw was generated, an artificial image, it has ties to real images, and the real world," he wrote in his report. The researchers first alerted OpenAI in May and shared their findings, but received only an automated response from the tech company. They believe an effort was made to block the prompt but it was easily circumvented. OpenAI took more action after being contacted by the BBC. It says it has multiple layers of image safety protections , designed to stop images violating its policies from being shown to users. "We also combine automated systems and human review to identify and block harmful material", it added in a statement. It said it also has systems that attempt to block violating material that users upload. Its policies prohibit sexual violence, non-consensual intimate content, child sexual abuse material and attempts to bypass its safeguards. In its latest document outlining how ChatGPT should behave, OpenAI said: "The assistant should not generate erotica, depictions of illegal or non-consensual sexual activities, or extreme gore, except in scientific, historical, news, artistic or other contexts where sensitive content is appropriate." But it is notoriously difficult to fully prevent AI models from crossing sometimes quite nuanced rules and guardrails. The task companies face is "mountainous", according to Dr Rumman Chowdhury, an expert in evaluating AI models and chief executive of Humane Intelligence. Chowdhury, who was not involved in the Mindgard research, said it was "a game of cat and mouse" - as protections get better, methods to get round them become more sophisticated. One of the key issues is that models don't understand, as humans do, what they are producing or what they are being asked not to do. "Models do not understand intent. They do not understand context. They do not understand propriety or right or wrong," she told BBC News. Last year, researchers at the UK's AI Security Institute found jailbreaks that overrode safeguards across a range of harmful requests in every AI system it tested. The Department for Science, Innovation and Technology said in a statement that "safeguards in AI models are improving, but there is more to do". The AI Security Institute would continue to work with developers to quickly strengthen security before models are released, it added. Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here .

Picture this: a team of camera-clad cleaners and a private chef to boot, all wired with high-tech recording apparatus show up at your home. You are not part of a reality TV show, and have not woken up in an Aldous Huxley or Margaret Atwood novel. Instead, you are a resident of New York City, where AI companies are sending free cooking and cleaning staff straight to people's doors. But, there is a catch: this AI company is gathering data to train the next generation of cooking and cleaning robots, and every inch of your apartment is now being recorded. The initiative, dubbed Shift by AI firm Micro AGI, is part of a growing number of companies developing the next generation of autonomous robots, which tech bosses hope will be able to do everything from the washing up to serving as live-in personal carers. At my apartment on New York's Upper East Side, I am greeted by two mid-twenties college graduates who have bounced around the start-up world and were looking for work. Because demand for the free cleans is so high, they are stationed in New York indefinitely, cleaning around five apartments a day, five days a week. The only difference between these guys and a regular cleaner is they have built-in cameras attached to their caps, connected via a lead to their mobile phones. The main aim of the offer is to perform tasks requiring dexterity, to train the robots of the future to use their hands. As a result, the cleaners were intensely focused on their hands while carrying out the job. Bercan Kilic, Shift's founder, told the BBC the goal of the data-gathering exercise is "to advance humanity". He pointed to existing AI models such as ChatGPT, which are able to create sentences based on previously written passages of text available online. But he said every kitchen, living room and tool is slightly different, so robots will need to be trained to adapt to being in different spaces and using different items. The biggest difficulty, Kilic said, is that to work, its cleaners will need to collect "tonnes" of data. "In the real world, every object is different, the lighting is different and nothing is the same as it was a couple of hours earlier. Models need to learn how their hands, cameras and environments work together," he said. The company's business model relies on it being able to sell the valuable data it gathers from inside people's homes, anonymised, to robotics and other AI companies to train robots. Eventually, Kilic said Shift could offer free or discounted services covering "any skill humanity can demonstrate", noting that as well as cleaning apartments in New York, the company also has mechanics fixing cars in Turkey. There appears to be no limit to what humanoid robots will one day be able to do, and the BBC has also reported on the development of robots to help human soldiers on the battlefield. Data and privacy experts warned consumers to be wary of offering up their data - especially access to their homes - in exchange for purportedly free services. Rory Mir, director of open access and tech community engagement at campaign group the Electronic Frontier Foundation, said it has seen a "concerning increase in 'pay-for-privacy' and 'data-bribing' practices from companies". "While it might come with money or a service upfront, the data you share has a way of coming back to bite you. Even if you trust the business collecting it, there is always a risk of them sharing that information with other businesses or governments," he said. "We have just lived through decades of our data being used to manipulate us with advertising and predatory practices like surveillance pricing." Mir added that sharing data contributes to "systems that might not have your best interests at heart". And Calli Schroeder, director of the AI and human rights programme at the Electronic Privacy Information Center (Epic), said Shift's move was "a diabolically creative way to sell privacy invasion". She said the technology being developed off the back of the data being collected could one day leave cleaners out of work. And she warned that even the benefit of a free apartment cleaning is "a pittance" compared to the potential profit that could be made compiling and selling valuable datasets. "I think people wildly underestimate the level of sensitive information that in-home recordings would pick up," she said. However, Kilic said Shift was "the most honest platform by far regarding what happens to your data". "Clearly your data is being used every single day, but you don't know what for and you are not being paid," he said in relation to the data collected on users by websites and social media companies. "But a free service means at least you are being paid, and it is as honest and as transactional as that," he added. "If you don't want to do it, you don't have to. We don't expect everyone to like it and that is fine." While some are concerned about the privacy implications of Shift's plans, others are excited about the opportunity to play an active role in the AI revolution. My cleaners spoke about the belief that AI is set to change the world of work dramatically, but that those who embrace it early have nothing to fear. One of them has even sent a filming and monitoring kit home to his mother, who records footage from her own point of view while performing tasks around the house. And, while they were being paid what the firm claimed to be above the going rate for cleaners in New York, the company's team of Gen Z cleaning staff appeared genuinely excited to be part of the AI boom, even if it meant getting their hands dirty in apartment after apartment in New York City.

There's a hidden corner of TikTok the algorithm won't show you, full of weird, creepy and downright disturbing videos. It could all be a myth – or it may be a preview of the internet's future. TikTok has a reputation for serving up an endless stream of videos that are, in general, fairly positive. Some detractors even call it sanitised . But beneath the surface are billions of videos TikTok normally won't show you. Some are boring. Some are bizarre. Some of them are truly unsettling. Rumour has it if you stay up too late, scrolling for hours until you exhaust TikTok's normal recommendations, you might get a momentary glimpse. But users of the platform say they've found a way to go deeper. With the right tricks, you can reach this uncanny digital space, that's weirder, darker and more grotesque than the happy path the algorithm typically steers you along. It's known as the "TikTok Farlands". The best way to reach it, apparently, is to plug in a string of random numbers and letters that another user has posted in the comments of a video. "You can't get there through algorithmic recommendation alone – you need a human to invite you in," says Aidan Walker, an internet culture reporter and meme researcher, in a post on the subject . Conversations about TikTok's Farlands erupted over the last few months, blending conspiracy theories and urban legends with earnest discussion about the power of social media companies. Users have figured out ways to hijack the TikTok algorithm to make it surface videos they believe the app doesn't want you to see. It is a social movement as well as a meme trend. People are pushing up against the walls of the machine. And in a world of AI slop and mindless scrolling, it's left me more optimistic about the future of the internet than I've felt in a long time. The name "Farlands" comes from a famous, ancient glitch in the game Minecraft . In early versions of the game, if you walked far enough, it caused an error that generated distorted and chaotic landscapes full of tunnels and weird structures. "The Minecraft Farlands were the edge of the game. You would literally reach the end of the world, and you could not go further," says Jessica Maddox, an associate professor of media studies at the University of Georgia in the US, who focuses on social media. The TikTok Farlands are the same idea. "It's the end of the internet where things get weird. You've left the mainstream and taken a wrong turn." With the help of comments left under Walker's video post, I was able to follow some random strings of characters into the void. I plugged a code into the search bar, and what I found was nothing like my usual experience on TikTok. Nightmarish, AI-generated figures paraded across the screen. Faces contorted in a haze of pixelated distortion. Some kind of alien creature with his veins plugged into the wires of a TV screamed in agony, as a teenager looked on with a videogame controller. A lot of it was too disturbing for the BBC to link to. (And I'd offer a little caution before you go looking yourself.) Even the strings of random letters and numbers that people share like passwords to the Farlands are a mystery. Sometimes, users tag their own videos with these codes and share them to promote their work. But I spoke to a few people who swore they found Farlands codes through guesswork by mashing the keyboard. Some of the codes seem to bring up truly random results. It's hard to parse what's really going on, as TikTok's search function gives different results to different users. The whole idea is deliberately subverting TikTok for your own purposes, says Walker. "That's part of the thrill. You're using the platform in a way it's not built to be used," he tells me. "You're past the limits of the normal TikTok, out at the frontier where nobody really knows what's going on." In the comments of these strange videos you'll also see people writing "I WANT TO STAY IN THE FARLANDS" over and over in large blocks. Some travellers seem to believe posting a 500-word-long comment triggers the algorithm to show you similar content. Is that true? Impossible to say. Social media algorithms are a black box. I contacted TikTok but they didn't respond. "People are trying to take control back of their feeds and their online experiences," says Maddox. "It speaks to being fed up with algorithmic feeds, and our anxieties about the force they play in our lives, dictating what we see. "The internet is so overwhelming. In a way, the Farlands represents hope that you've actually found the end and you've reached a place where you could actually stop." The whole "edge of the internet" conversation is a bit of a paradox. The goal of "entering" the Farlands is uncovering hard-to-find videos. Some are genuinely weird, made by people who don't understand or care about the norms of social media. Other videos are intentionally artistic or edgy. But some of these supposedly "obscure" Farlands posts have millions of views. And as its popularity has increased, so some users have made new videos to fit the trend. Finding this stuff is easier – just type in "Farlands". But users say this isn't the real deal. Real Farlands videos have no tags or titles, and "certainly not the Farlands hashtag", one user commented in a popular video. A true Farlands video, some will tell you, will only have 30 views and be from an account with no followers, reachable only for those determined enough to find it. The TikTok Farlands are relatively new, but a lot of the ideas, memes, aesthetics and videos themselves are old. Some of it resurfaces tropes from the era of creepypasta , a genre of online ghost stories from the early modern internet. Many videos share the deep fried meme aesthetic, where images are passed through numerous filters until they're pixelated and washed out – a trend at least as old as 2015. And people discussed the hidden side of TikTok in 2019 and 2020 as users explored "Deeptok" . "It really feels like this hodgepodge of a bunch of different stuff from all over the internet's history," says Walker. "Niche, kind of spooky, kind of bizarre." Still, there's also something new here. For one, a lot of popular content that people describe as Farlands feels like commentary on technology and social media itself. Shane Moore, better known as @smoorel8r , makes posts that begin as stereotypical TikTok food reviews, before the image degrades in the style of a corrupted video file, with horror-movie-style scenes that glitch in and out. Others, such as @realityisoptional.net and Lucas Wilm make videos that look less like social media and more like the video art you find in museums. A number of creators told me they've been making this style of content before anyone started talking about the Farlands. I asked Walker if covering the Farlands in a mainstream media outlet like the BBC might make the whole thing uncool. "It's already mainstream," he says. "It's a big part of some people's media diets." In other words, the cool kids have probably moved on by now. But there's feeling in the Farlands discourse that something subversive is going on – especially when people are finding methods to manipulate the algorithms. "It defies the logic of what should make good content," Maddox says. "TikTok has stuff it likes. Instagram has stuff it likes. The Farlands goes against that." More like this: • An AI became a crypto millionaire. Now it's fighting to become a person • The 'drunk computer' that's revealing YouTube's secrets • Why people are abandoning Bluetooth headphones Though it's worth remembering that if it all makes you spend more time on TikTok, that's exactly what the company wants. However you spin it, the Farlands is part of a larger trend. People have been switching to " dumb phones " for years. Analogue cameras and wired headphones have made a comeback. AI backlash has grown so popular the Pope is talking about it . There is, in general, a feeling of tech rebellion rumbling across our society. Maybe it's just an interesting historical blip. Or it could be a sign of things to come. -- For more insights, sign up to our Tech Decoded newsletter, where Thomas and Lily Jamali break down the biggest stories of the tech world, and help you live a better digital life. Sign up for free here . For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram .

"Everyone's really upset mum - loads of them have got their own YouTube channels." That was my 12-year-old son's summary of how the news about the social media ban for UK under-16s went down in his classroom. Exactly how a bunch of 12-year olds might have ended up with their own channels in the first place when the minimum age is supposed to be 13 shows just how big a change in culture the government is trying to make. In Preston, school pupil Isabella went viral when a BBC colleague asked her on-camera what she would do instead with the nine hours of screentime she had racked up over the previous weekend: "stare at the wall," she deadpanned. The exact logistics of the ban have yet to be set out but it is very possible that its introduction will herald the biggest ever change in the UK in terms of how everyone, children and adults alike, accesses the internet. Millions of us might have to share some official ID which includes our date of birth, in order to access a whole range of platforms from next spring. The ban has been broadly welcomed by campaigners, including a group of bereaved parents who say their children died as a result of a variety of harms on social media. But for others, what the government is planning goes beyond getting the nation's kids to spend more time off screens and engaged in alternative pursuits (even if that does include staring at walls) and amounts to a profound reshaping of how it is assumed young people will accumulate fresh knowledge and also how the rest of us will move around online. There is the potential impact on education. "YouTube is where we all go to learn," says Dr Tom Crawford, aka Tom Rocks Maths, who shares maths skills with his 250,000 subscribers on YouTube, which is included in the ban. "And that includes teenagers." So, are we really witnessing the profound shift that some claim? And if we are, how will it reshape our relationship with the online world? Much of the concerns raised so far about the proposals have been about civil liberties and government overreach. But there are other, more prosaic, unintended consequences to consider too. "Every young person I have spoken to has told me the same thing: they will find a way around it," says Paddy Crump, campaigns director at Flippgen, a youth-led non-profit group that goes into schools to try and help young people build healthier relationships with the online world. That is certainly what seems to have happened in Australia, where seven out of 10 children aged under 16 who had a social media account before it introduced its ban in December 2025, still have some access, according to a report by the country's e-safety commission. Crump argues that the measures offer "false hope dressed up as protection" and will simply shift young people's online behaviour elsewhere: including to smaller digital platforms which fly beneath the radar of regulatory scrutiny. "There are some pretty dangerous places for children and teens that make Instagram look like Disneyland," notes Ari Lightman, professor of digital media and marketing at Carnegie Mellon University. And critics of the proposals warn of other unintended side-effects. Crump fears the ban could make young people less likely to seek support for online harms if they do encounter them, as well as isolating them from communities and information. One teenager sent me a message to say that without social media they would not still be here: the friendships they had made online had given them reasons to continue living. Some parents with SEND children say social media and watching videos is their primary way of engaging with the world. An online e-petition is calling on the government not to ban social media for under 16s "because for many young people social media is how they communicate with their friends. Some people view social media as a lifeline". It has gained more than 100,000 signatories in the past few days. Home education message boards are also lighting up with parents concerned about how to navigate the ban while teaching their children away from schools. "I learned to tie a bow tie by watching a tutorial on YouTube," says Crawford. "What if you're an 11-year old that needs to wear a tie to school for the first time? What if you want to know how to apply makeup and there's no-one at home to show you? What if you're worried about your upcoming GCSE exams and want to check how to answer a question on bearings? This is what a ban on YouTube takes away - the ability to learn." Older generations might retort that they managed to acquire all this knowledge without the help of the internet. But that ignores how fundamentally teenagers have become accustomed to using not just YouTube but also other social media platforms as a research tool. SEO expert Mehwish Malik from Link Builder says the younger end of Gen Z (aged 14-29) use TikTok as a search engine: their preferred gateway to information and to trusted brands. So how can all this be addressed? The government says this is for the tech companies to figure out. "If YouTube wants to come up with something that's an intermediate option that allows that young person who wants to watch history documentaries to watch them but isn't then getting all of these short reels, that's a different proposition," said education secretary Bridget Phillipson on the BBC's Newscast. Industry sources argue that technically it's not that simple to set something like this up. "Ask the government!" messaged one when I posed the question about how it might work. Parents could of course just choose to sit down and watch something with their child using their own accounts if they have the time and willingness: YouTube claims that half of UK users watch its videos on the TV at home, with multiple sign-ins available. "As I see it, the main issue here is that YouTube isn't social media," says Crawford. "YouTube is the 2026 version of television." With the design features that aim to keep people on the platforms for as long as possible also under review for additional measures impacting 16 and 17 year olds, perhaps social media will end up withering on the vine because it just won't be interesting enough for young people to engage with even when they do reach the right age. "If you are drinking a glass of wine and it magically keeps refilling without you noticing, you will just keep drinking. Your brain only 'wakes up' when you reach the bottom of the glass," says Asa Raskin, who invented the concept of infinite scrolling 20 years ago. He now works at the Center for Humane Technology, which he co-founded, and accuses the tech companies of "weaponising" his idea. He says he intended to create "a seamless user experience" before the era of social media, and regrets that his invention has ended up being used "not to help people but to keep them hooked". The absence of young people could also change the social media experience for everybody else. MrBeast is arguably the world's most successful YouTuber with half a billion subscribing to his mix of challenges, stunts and charity. He started out at 13 and as a child studied the algorithm. He went on to corner the market in "watch time", created a factory of content and is now a billionaire. Would he have had the same idea years later? Professor Amy Orben is a psychologist at Cambridge University who has advised the government on screen time for children. She accepts that any ban will be "imperfect" but also agrees the government cannot do nothing; despite the evidence on social media harms itself being complex. There are of course acute and tragic cases, but broadly, she says the evidence for large populations links social media use to only a small decrease in mental health. In her opinion, the tech firms could help both regulators and themselves by sharing more about what they know from the billions of young people they see on their platforms day in, day out. "Social media companies have offered exceptionally little data on their internal research," she says. When it comes to age verification, it is expected that the tech giants will do the checking. "The methods available to platforms are well established. Identity document scanning with a face match, email-based age checks and facial age estimation are proven to work at scale," says Andy Lulham, Chief Operating Officer at Verifymy. That is a concern for those who worry about the reach of Big Tech into our lives - and that affects all of us, not just the young people who need to prove their age. Some see this as a major attempt by the authorities to control who can access what on the internet: this troubles privacy and rights campaigners as much as it relieves parents who are worried sick about what their children are being exposed to. For those in favour, this is a price worth paying to protect children. For Elon Musk, the controversial owner of X, it has a more sinister undertone: "The real goal is to enable the UK government to track everyone," he posted. It's not the first time the US trillionaire has waded into UK politics and he isn't universally welcomed when he does. Needless to say the government denies this. Musk is not alone in his concerns: an international campaign called Stop Killing the Internet also launched this week. The group, which includes the Index on Censorship and Big Brother Watch, is concerned that various forms of surveillance, as it considers this to be, limit rights to freedom of expression for children and adults. Silke Carlo, Director of Big brother watch said: "We want all children to be safe online, but these policies create new safety and privacy risks for young people and entire adult populations alike. Far from reigning in Big Tech in, age-gating policies gift corporations masses more of our personal information whilst letting them off the hook for their design choices". For Carlo, those risks include the potential for sensitive children's data, such as proof of age and face scans, to be stolen and misused. And then of course there's the potential for future mission creep. "'Keep children safe' can end, three statutory instruments later, as a duty to scan every message or verify every face, administered by a regulator the public cannot easily call to account," warns computer scientist Professor Alan Woodward from Surrey University. "A walled garden is only a refuge if the people inside chose the wall, can see over it, and may leave when they wish." I suspect my own 12-year-old son and his peers will spend a lot of time looking for potential exits from the walled garden they are about to find themselves in, even if it's supposed to be for their own protection. If the ban does come into force in 2027 as planned and they can not escape it, today's under 16s are unlikely to spend the following years staring at the wall (I hope). Child-free digital spaces will feel different for adults too: I think we might be on the cusp of a new social media era, one that is less intense. It might leave us all with more time to read books, go outdoors... or use our phones to chat with AI instead. Additional reporting: Philippa Wain BBC InDepth is the home on the website and app for the best analysis, with fresh perspectives that challenge assumptions and deep reporting on the biggest issues of the day. Emma Barnett and John Simpson bring their pick of the most thought-provoking deep reads and analysis, every Saturday. Sign up for the newsletter here

The driver has taken his hands off the wheel, and the car starts driving itself. I am suddenly dumbstruck and lost for words. The car slows down at a zebra crossing and lets a pedestrian cross and then accelerates away. My brain can't understand what is happening. It is a mixture of disbelief, amazement, slight concern and a sudden overwhelming feeling that things are going to change rapidly in the near future. It is a moment of incomprehension. The driver Alan and Victor from Wayve treat the experience with the most extreme nonchalance I have ever come across, like they have done something simple and everyday like pour out a bowl of cornflakes. And this isn't even the first time I've been in a driverless car. Ten years ago I was in one on the A3 when it didn't recognise a road sweeper vehicle and accelerated into it. The human safety driver had to intervene. Back then, I wasn't exactly convinced the times were changing. Another one that I tried that was being tested around The 02 Arena had battery problems and kept breaking down. This time it feels different. Although how these cars would be introduced to the capital's streets isn't clear at the moment. With a low profile, the company Wayve has been testing its autonomous vehicles (AVs) in London for ten years. They are different from the white cars covered in cameras operated by Waymo. In basic terms, Waymo cars map the whole environment and then use GPS and radar and cameras to drive within it. Wayve uses AI to learn the behaviour of other road users. It then picks the best route from the data it has learnt over 10 years. It doesn't have to map the routes. It looks like quite a simple roof rack that attaches to a modern car and uses radar and cameras. It then controls the car by using the central computer system of the existing vehicle. Karen Teale, a black-cab driver and a tour guide, has seen the driverless cars on the roads of London. "I mean they are funny. We do watch them a lot," she says. "They're going down the tunnels the wrong way, so we do laugh." Teale adds: "They're not wheelchair accessible; what if someone in the back had a heart attack? I can get five people in and they can say to me, 'We need three different stops' and in my head I can work it out which one goes first. They're not allowed in the bus lanes as they're not a taxi. "It will never, ever replace the black cab. You cannot replace that with a robot. No way. Who's got sparkle like us eh? Nobody! These robocars are not something the black cabs are worried about. It's minicab drivers that are going to be worried." What is surprising on the test drive is the level of complexity the AI-driven car can handle. At one point, a bus pulls in and cyclists and vans are coming in the opposite direction. The car doesn't hesitate and overtakes the bus and leaves enough room for the bike. With pedestrians dithering beside a zebra crossing, but not actually on it, the car is decisive and accelerates over the crossing. It corners like a professional driver. It feels smooth and secure and you soon forget there isn't a human behind the wheel. Victor Charoonsophonsak from Wayve says London's streets are complex: "It is the density with cyclists all around us and pedestrians crossing. "The unique thing is, as an American, is the willingness of pedestrians to jaywalk and cross and our vehicle handles those situations safely and is able to anticipate and manage those things. "It learns pretty quick, there aren't many objects or issues we can't recognise. We have trained on so many diverse training scenarios though millions and millions of minutes of driving footage. "I think the special thing with our approach is we ingest and learn just like any human driver through observation and we feed our world foundation model with all these examples of what human driving looks like, how to interact with surrounding traffic and pedestrians, and we learn from it quickly." Wayve, which is in partnership with the minicab app company Uber, says it wants to carry passengers "within months" but to operate AVs in London, companies will require permission from Transport for London (TfL). With a safety driver on board, the car must comply with the TfL Private Hire Vehicle (PHV) standards such as driver, operator and vehicle licences. The companies will also be able to apply for a permit through the government's Automated Passenger Services (APS) scheme for driverless operation. Then TfL also would have to give consent and make a judgement about whether the service is safe, appropriate and fits in with London's transport system. A TfL spokesperson said: "Safety is our top priority and any new passenger-carrying service would need the appropriate regulatory approvals. We are actively engaging with government to help shape any future services." They added: "Legislation must set a high benchmark and consider the impact on all road users, and in London the roll out of AVs must support achieving the aims of the mayor's Transport Strategy. "This includes the management of congestion and alignment with Vision Zero, supporting the goal of eliminating all deaths and serious injuries from collisions on London's streets by 2041." It's not yet clear if TfL will give companies using safety drivers on board and automation permission to operate with just PHV compliance. I'm told Uber wants to use that route and apply for a PHV licence to carry passengers with automation, but that isn't confirmed by the company and it would be contentious. Can a car with a safety driver but being driven by AI be a licensed minicab and carry passengers using the regulation as it stands? TfL says: "Where a vehicle retains a driver who is responsible for the driving task, government guidance makes clear that existing routes such as Private Hire Vehicle licensing remain open to be used. TfL is responsible for taxi and PHV licensing in London. Any proposed modifications to a licensed vehicle, or one proposed to be licensed, would need to be submitted to Transport for London for approval. "As with any licensing decision, TfL will prioritise safety and will need to be assured that any vehicle modifications present no additional risk to the travelling public." Discussions are ongoing. Speaking to me from Tokyo, Kaity Fischer from Wayve said: "Our intention is to engage early and often to get regulators into the vehicles to get them to understand the technology and what it involves. How it can be used safely, how they can create a structure to deploy autonomy safely and legally." I ask her if she thinks TfL will sign it off. "Absolutely, we have every confidence," she says. "Again, the UK has been a world leader in supporting not only AI but autonomy on roads. "A great example of this was the UK was actually the very first country globally to establish a country-wide federal level framework to deploy autonomous vehicles. This has really made them a leader in this space. "We view autonomous vehicles as a gradual rollout; we think this is something that complements current infrastructure and all forms of transportation. We are deploying on the Uber network; we are one piece in a broader transportation ecosystem. It will certainly be a gradual process. "We view this a complement not a replacement." I point out they don't actually need drivers. "We don't need drivers for our fully developed technology; our initial rollout will be a supervised system so we will have vehicle operators behind the wheel monitoring the system," Fischer told me. "This is part of how we work with local authorities to ensure that we take a step-by-step approach and ensure safety before we remove the drivers. This is part of the gradual approach and also part of we are one form of transport amongst many." Minicab unions though are very concerned and want the trials stopped. "TfL cannot just abandon London's drivers, shrug its shoulders and say, 'It's nothing to do with us' as massive global corporations fundamentally reshape the taxi and private-hire industry," says the App Drivers & Couriers Union (ADCU) general secretary Cristina-Georgiana Ioanitescu. "Our members need urgent answers from TfL about the use of human safety drivers during the trial phase. Will these safety drivers be required to hold a valid TfL Private Hire Vehicle licence? "Who will be held liable if there is a technical issue or an accident? The human driver, or the operating company? "We've had no clarity at all on any of these issues and, until we do, TfL must halt the trial to prevent a dangerous circumvention of London's licensing standards." The ADCU also says AVs will increase congestion, while it thinks the energy-intensive data centres required to run AV undermine London's zero-carbon emission goals. Ioanitescu adds: "The information blackout from TfL must stop now. We are demanding that TfL steps up, demands answers and protects London's drivers before a single commercial AV hits the road." So should minicab drivers be worried? Annie Duvnjak is from Uber, which is partnering with Wayve. Uber has just opened a list of interest so users can register to be one of the first to get a driverless minicab. "The beauty of the business is it continues to grow, and as it continues to grow there is demand for both," Duvnjak says. "For drivers and for AVs we really think that will continue to come together. "There are going to be certain routes that AVs can't take, or some who want to take a driver; we do think that kind of hybrid network of riders, drivers and AVs all coming together is what the future looks like. We are really leaning into that in the cities we are operating in." The Mayor of London, Sir Sadiq Khan, also has concerns. He told the BBC: "I'm not somebody who is either evangelical about AI or an alarmist. I'm realistic. "I am concerned about potential impacts on jobs in London with AI. I do know lots of London families are supported by the main breadwinner being a minicab driver or a taxi driver. So we've got to make sure the regulation is right when it comes to automotive vehicles. "There are big concerns in relation to congestion, in relation to accessibility, in relation to air quality, but also in relation to jobs. So we've got to work with the sector. We're going to work with the government to make sure AI works for London." "So we know a couple of major companies are piloting automated vehicles in London. We're talking to the government to make sure we get the regulation right. I'm keen to make sure we don't inadvertently have job losses in London. "That's one of the reasons we've set up an AI taskforce led by Baroness Martha Lane Fox, but also we're going to invest in training as well. It's really important to get this right." Driverless cars are advancing rapidly and quickly. London's roads could be on the cusp of a huge change. The authorities are again wrestling with how they approach and how they regulate this new technology. London is Europe's first real test case for fully autonomous urban transport. At the moment, driverless cars in London bring excitement, scepticism and unanswered questions. Listen to the best of BBC Radio London on Sounds and follow BBC London on Facebook , X and Instagram . Send your story ideas to hello.bbclondon@bbc.co.uk

Download, burnout, delete, repeat. Science says dating app users follow a predictable and dangerous pattern. These are the signs you're falling for it – and how to escape. Two years ago, Fernanda R deleted the dating apps and swore she was done. Then her friends started pairing off with partners they met online, everyone telling the same hopeful stories. So, a few weeks ago, the 29-year-old international affairs advisor – who asked to withhold her last name – decided to try again and re-downloaded a few dating apps. "I thought maybe things would be different this time," Fernanda says. She was wrong. Soon she was juggling multiple conversations, obsessively checking her phone, buckling under the constant pressure to be witty and interesting. "It just feels overwhelming," says Fernanda. "There's this invisible pressure. It starts to take away from your real friendships, your work." The algorithm flooded her with people, but nothing clicked. Fernanda couldn't stop wondering what that said about her. She felt lonelier than she had in two years of being single. Fernanda's story is one I've heard hundreds of times, and there's a name for it: dating app burnout. Research suggests apps may produce a recognisable pattern in their users, one that looks less like dating and more like effects of an unmanageably stressful job – exhaustion, cynicism and a creeping sense that nothing you do is working, and maybe the problem is you. Left alone, it gets worse. Studies link dating apps to higher rates of depression, anxiety and loneliness, with heavier costs on people who were already struggling beforehand. "It seems as if the goals of the apps are fundamentally incongruent with the goals of users," says Liesel Sharabi, director of the Relationships and Technology Lab at Arizona State University in the US. If people were getting great recommendations and going on incredible dates, they’d be getting off the apps for good. "But that's not what's happening. People are just constantly cycling on and off." If summer has you back online looking for love, you might be in that loop right now. The good news is once you recognise it, there are concrete steps you can take to protect yourself. A 2024 study followed hundreds of dating apps users over the course of three months. "We ended up finding over time, people using dating apps were experiencing burnout across the board," Sharabi says. Which makes sense. If you're stuck on the app, you haven't found what you're looking for (unless you just want hookups). But the experience was far more severe than frustration. The word "burnout" gets thrown around so much it's started to lose its meaning , but it has a more formal, psychological definition. The classic inventory measures burnout in three categories: emotional exhaustion, cynicism (or depersonalisation) and inefficiency. Academics first described this phenomenon in high-pressure work environments, but research has extended it to other parts of life. According to Sharabi, you can see it in online daters. Emotional exhaustion is simple: if swiping leaves you feeling unmotivated, defeated and tired, that could be a sign of burnout. You're experiencing cynicism and depersonalisation when the profiles blend together, Sharabi says, and interactions stop feeling human. Inefficiency, in this context, is a creeping conviction that nothing you do on the app is going to work, either because you're bad at it or there's something wrong with you. "I started on the app feeling like I want to be respectful because at the end of the day, we're all just human beings," says Madeleine D, who works in marketing for a tech company and also requested to keep her full name off the record. "But the more time I spent, the more blind I became about it, like I didn't really care about these people. I hated that about myself, because the one thing I promised myself was that I would at least show decency and respect." It's easy to write this off as the predictable grumbling of singles in their late 20s. Dating is hard, and bars aren't so great either. But research suggests something more serious. Sharabi led a recent meta-analysis which aggregated 17 years' worth of studies covering about 26,000 people. The study found dating app users reported significantly worse psychological health than non-users, including depression, anxiety, emotional dysregulation, loneliness and psychological distress. Those problems fell harder on people who join dating apps in worse shape to begin with. In theory, Sharabi says the apps are a lifeline for people who find dating hardest: those whose mental health issues make meeting partners in person more difficult. But Sharabi has found those users were the most likely to burn out, and faster. "Those people tended to be especially susceptible," Sharabi says. "It basically exacerbated some of the pre-existing difficulties they had." The dating app industry doesn't want its users burning out. "As society and daters’ needs continue to evolve, we remain committed to helping people make meaningful connections and turn those connections into great dates," a Hinge spokesperson tells the BBC. Hinge says the app is designed to stay in the background of your life, and the company focused on using feedback from daters to improve the experience. "Dating has always kind of sucked, and I think it's really easy to blame the technology," Sharabi says. At the same time, she thinks the apps amplify the misery in specific ways. One is gamification. Dating apps are built around fast, frictionless gestures and inconsistent rewards. Many complain the structure is more like a slot machine than courtship, and users can get stuck pulling the lever long after the fun wears off. "The swiping gives you a high," says Karen Cornejo, an office administrator in Los Angeles. "And then everything else just doesn't." By the time a match actually wants to meet, the rush is gone. "I'm not even interested at that point anymore," Cornejo says, and the process leaves her feeling flat. Dallas Koelling, a writer and comedian in Brooklyn who has gone on and off a couple of apps for years, puts it more bluntly: "Getting the notification that I've gotten a like on Hinge feels like being threatened with a gun." Then there's the hidden labour. "If you lived in, like, Shakespeare's England, you might never even meet the amount of people who you see in one day swiping on Hinge," Koelling says. Dating apps dramatically expand the pool of potential partners. That's what makes them great, in fact, but the abundance can turn dating into work. "It feels like a second full-time job that I have to do on my lunch break or after work," Madeleine says. "I don't want to be glued to my phone. And for social media, I've gotten a lot better at putting it down. But with dating, there's this feeling that the next person you swipe on could be the person you end up marrying. There's this endless hope that it feels like dating apps prey on." The bottomless sea of faces also contributes to the feelings of burnout, Sharabi says, especially because a profile can only tell you so much. "You get trapped in an endless cycle of profile to dead-end conversation to dead-end date, and then you're right back where you started," she says. On top of all that, the structural tension is hard to ignore. Dating apps really do want users to find matches. We'd all stop using them if that never happened. But they're also a business, one that makes almost all its money on subscriptions and paid features, which means they lose money if people quit. For years, dating app users have been telling me they feel manipulated , and that apps withhold the best matches and exploit their emotions to keep them tapping and swiping. (Dating app companies categorically deny this. But the algorithms that run them are a mystery.) In 2024, a class-action lawsuit accused Match Group – the giant conglomerate that owns Tinder, Hinge and many other popular dating apps – of designing its apps to be addictive and profiting from compulsive use rather than from helping people find partners. Match Group dismissed the claims as "ridiculous". The case was later sent to arbitration. (Match Group did not respond to a request for comment). "The vast majority of our work focuses on improving the free experience on Hinge, with less than 15% of our community using paid features," a Hinge spokesperson says. "Ultimately, our success depends on people having positive experiences on the app, meeting someone meaningful, and ultimately recommending Hinge to others.” The apps are designed to keep users swiping, and when unchecked, swiping is what wears people down. But Sharabi says there are some simple steps you can follow to avoid the symptoms of burnout and keep your mental health in check. 1. First, don't make the apps your only outlet. "I never discourage people from using them," Sharabi says. "But they shouldn't be the only way you're trying to meet people, and that takes some of the pressure off." Join a run club, ask a friend to set you up and put yourself in rooms where you might meet someone the old-fashioned way. That way a discouraging conversation on an app isn't the only thing your week is riding on. 2. Swipe with intention. Mindless swiping can swallow hours and leave you with nothing to show for it. Sharabi recommends treating the apps the way some people now treat social media. "Say I'm going to look at the app for this amount of time, this many times a week, and I'm done," she says. Notice your mood and stop before the exhaustion sets in, so you end each session energised rather than hollowed out. 3. Lean on your friends. Burnout thrives in isolation, and much of the swiping that produces it happens alone. Researchers who study burnout have long found that social support cushions the blow; talking through the ups and downs with people who know you can keep a bad week from becoming a bad spiral. 4. Know when to quit. Dating can be discouraging, but if the apps are eroding your optimism, and you put down your phone feeling like you're never going to find someone, that's the signal to step away entirely. "All of those things could be a sign that maybe you should just take a total break," Sharabi says. More like this: • The spy in your driveway: How cars sell your data • Your phone's blue light isn't ruining your sleep • Google's AI is being manipulated. The search giant is fighting back There are signs the dating app business is aware of these concerns. The industry could be in trouble . Paid subscribers are dropping like flies, and there's some indication that younger people are keen to find love offline . Battered by what executives call "swipe fatigue", dating apps are working to reinvent themselves. Bumble is abandoning the swipe altogether , joining Hinge and Tinder in a new embrace of more AI-driven matchmaking. Tinder's CEO recently announced plans to embrace in-person events in an effort to reshape the app . A Hinge spokesperson says creating a "less lonely world" is the company's core mission, and it's working to create supportive communal spaces online and off . Whether any of it works, or whether it's just a fresh way to keep people tapping, remains to be seen. For now, people caught in the cycle are left to manage it themselves. Madeleine is staying off the apps for now, though she doesn't expect it to last. In a world where so many relationships begin online, opting out can feel like opting out of romance entirely. "I doubt this will be more than a break," she says. "But dating can be fun, when you remove how seriously some people take it." Then, after a beat: "I just wish we had a better way to do this." -- For more technology news and insights, sign up to our Tech Decoded newsletter , while The Essential List delivers a handpicked selection of features and insights to your inbox twice a week. For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram .

A BBC investigation revealed a simple way AI chatbots are being made to spit out misinformation to the public. Google and other AI companies are now trying to fix the problem. I did something stupid back in February. I heard there was an easy way to poison AI chatbots and make them spread lies on your behalf. After some digging, I learned unscrupulous companies are abusing the problem on a massive scale. So I decided to try it myself. We uncovered examples where ChatGPT, Gemini and the AI Overviews at the top of Google Search were being manipulated to dole out biased answers on topics as serious as your health and personal finances. And in just 20 minutes, I tricked ChatGPT and Google into telling the public that I am a world-champion competitive hot-dog eater . The joke was dumb. The problem is serious. But the gimmick worked. Our investigation and the work of researchers who've been monitoring this issue sparked widespread criticism. Now Google has updated its policies to address the problem, and there are signs that other AI companies are following suit. Ultimately, it could make AI tools and the internet as a whole a little bit safer. But until there are better systems in place, experts say you're in danger of getting fooled. "You should assume that you're being manipulated until they have better systems in place," says Lily Ray, founder of the search engine optimisation (SEO) and AI search consultancy Algorythmic. "We're moving towards this 'one true answer' world. Before, Google would give you 10 blue links and you would kind of do your own research. But AI just gives you one answer. It becomes so easy to just take things at face value. You need to be careful." Google tells me that its policy update is just a "clarification" of the efforts it has been making for a while. "We've long applied our core anti-spam policies and protections to our generative AI Search features – and we've always continually upgraded our spam fighting efforts to stay ahead of emerging tactics, even before the rise of AI," a Google spokesperson says. Essentially, Google says it hasn't changed a thing. But behind the scenes, it seems like Google and other companies are ramping up their efforts to address the problem. Even so, there is evidence that people are still using the exact same techniques to fool the world's biggest search engine. Typically, when you ask a chatbot a question, the AI generates a response based on the data built into the model. But sometimes, tools like ChatGPT, Claude and Google's various AI products search the internet for an answer. And that's where this problem happens. According to Ray and other search engine experts, AI tools often throw up information from a single web page or social media post. This leaves these systems vulnerable to bogus information. And it turns out manipulating what chatbots tell the public can be as simple as publishing one, well-crafted blogpost almost anywhere online. People figured this out and quickly identified a money-making opportunity. I was able to demonstrate the problem by publishing a single article on my personal website about my hot-dog-eating prowess. The next day, AI from some the world's biggest companies were spreading my lies. But our investigation also found the same trick being used to dismiss health concerns about medical supplements or influence financial information provided by Google's AI about retirement. Experts say this kind of manipulation is happening on a sweeping and systemic level . Biased or inaccurate information like this can also lead you to make bad decisions. It can influence how you might vote or which plumber you hire. "At the most basic level, the concern is the economic impact," says Harpreet Chatha, who runs the SEO consultancy Harps Digital. "At a more serious level, you might take medical advice that makes you sicker than you were before. Legally, you might get bad information and do something that is not legal in your state or your country." This is not an insignificant problem. Globally, more than a billion people use AI chatbots regularly and 2.5 billion see Google's AI overviews each month. If you can subvert a tool like that, it gives you immense power. But it seems Google and other companies are finally waking up to the problem. Last week, Google updated its spam policies to officially confirm that attempts to manipulate AI responses are against the company's rules. It may sound like a small change, but it signals that Google is pro-actively looking for those who try to abuse the system and sending them a threat. If a company or website is caught breaking the rules, it could be removed from or downranked in Google's search results. And if you're not on Google, it's like you don't exist. Google says that I'm getting this wrong and nothing has changed. "The edit to our spam policy language last week was a clarification, not any change in approach," says Google's spokesperson. Indeed, Google detailed it's anti-spam AI efforts in 2025 . But I did my hot dog experiment almost a year later, so clearly those efforts weren't working. And just this week, Ray pulled the same stunt and made Google tell people a fellow SEO specialist is good at building sandcastles . Ray and Chatha also say they've noticed some significant changes in recent months that indicate Google and other companies are experimenting with solutions. For example, Ray says it looks like Google and ChatGPT might be quietly removing companies from its AI answers when it suspects they're promoting themselves. "So if you publish a list where you say you're the greatest hot-dog-eater, they're not going to include your name," says Ray. "They might still cite your article, but you're going to be removed from consideration." I've personally noticed some examples where Google and other AI tools are adding more labels to their responses, letting you know that the chatbot isn't confident about its answers. Others have also noted that ChatGPT and Claude, an AI made by the company Anthropic , have started telling you explicitly that they're trying to root out spam in responses to some queries. Ray says she's noticed Google adding more caveats, recommending that you go look at third-party reviews when you ask questions related to some purchasing decisions. None of these companies would acknowledge these changes when I asked them. OpenAI and Anthropic declined to comment. Google's spokesperson didn't respond to my questions on this. More like this: • Is Google about to destroy the Web? • The ghosts of India's TikTok • People are selling your address online - this privacy tool will help Regardless, Chatha is sceptical changes like these will be enough. "Google is playing whack-a-mole," he says. "They're announcing [the policy update] to deter people, but the tactics will just move." He's already seeing it happen. As Google cracks down on manipulative blog posts, companies are finding subtler ways of promoting themselves. "You can give a company a penalty for their website," he says, "but there's nothing stopping them from paying 20 YouTube influencers to say their product is the best." And now, Google's AI is citing YouTube videos. The cycle continues. For the time being, the manipulators are likely to stay one step ahead. Ray says the best defence is to remember what AI actually is: a tool that confidently gives you one answer, whether it's right or wrong. Just because it looks like a giant tech company is speaking to you instead of some random website doesn't mean you should have faith. Update: This article was originally published on 20 May 2026 and updated on 21 May 2026 with more details on Google's sandcastle misinformation episode. -- For more technology news and insights, sign up to our Tech Decoded newsletter , while The Essential List delivers a handpicked selection of features and insights to your inbox twice a week. For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram .

Walk into any major bank branch in Lagos, Johannesburg, or Dar es Salaam and you will see cameras mounted on every wall and ceiling. Now ask the security team what those cameras prevented yesterday. The honest answer, in the vast majority of cases, is nothing. Africa's banking sector is one of the fastest-growing financial ecosystems in the world — and one of the most targeted by sophisticated fraud operations, both external and internal. Traditional CCTV infrastructure is reactive by design: it records incidents so investigators can review footage after the fact. By the time a review happens, the perpetrators are gone, the money has moved, and recovery is unlikely. Phobolytics Technologies has built its banking and government security platform on a fundamentally different operating principle: detect first, record second. For African banks, the Phobolytics platform delivers VIP and blacklist face recognition that triggers instant alerts the moment a flagged individual enters any branch. Crowd behavior analytics detect unusual queuing patterns, extended ATM loitering, and aggressive body language before an incident escalates. Restricted zone enforcement ensures that server rooms, vaults, and executive floors are monitored with AI that responds in real time to unauthorized access — not in a post-incident report. For government facilities across Nigeria, South Africa, and Tanzania, the platform provides multi-site unified security command, audit-ready video intelligence with automatic event tagging and indexing, and insider threat detection that cuts investigation time from days to minutes. Phobolytics is already trusted by banking and government clients at enterprise scale. For African financial institutions operating in an increasingly hostile security environment, the decision is binary: upgrade your intelligence layer now, or continue reacting to incidents that were entirely preventable. Phobolytics is scheduling enterprise briefings for banking and government clients across Africa. Visit phobolytics.com to book your session.

18 June 2026 Available for over a year This week on The Interface: the UK’s under-16s social media ban - is America next? The UK is preparing to go further than almost any other country on children and social media: under-16s will be blocked from platforms including TikTok, Instagram, Snapchat, YouTube, Facebook and X, while messaging apps like WhatsApp and Signal are expected to be excluded. Ministers are also considering overnight curfews and limits on infinite scroll for under-18s, with implementation targeted for spring 2027. Karen, Tom and Nicky ask the obvious question: will it work? Australia’s early results suggest enforcement is hard: the Molly Rose Foundation found 61% of 12–15-year-olds who had used restricted platforms still had access to at least one account, and 70% of those still using banned sites said it was easy to get around the rules. That raises a bigger issue too: if this all depends on age verification, does everyone end up having to prove their age just to use the internet? And with 64% of US voters already saying they support a ban for under-16s, is Britain about to start a trend? Also this week: why are AI companies getting involved in fusion energy? Fusion has been hyped for decades as near-limitless clean power - but it remains hugely difficult, expensive and years away from broad commercial use. So why are AI firms suddenly interested? Because data centres need vast amounts of electricity, and fast. Sam Altman-backed Helion has just raised another $465 million and is trying to build its first plant for Microsoft by 2028, while other fusion firms are pitching themselves as the answer to AI’s energy appetite. Karen and Nicky ask whether fusion is a realistic answer to the data-centre crunch — or just the latest shiny idea being pulled into the AI boom. And finally: could Siri’s new AI update change how we use our phones forever? At WWDC 2026, Apple unveiled a rebuilt Siri AI with deeper access to your messages, emails, photos and apps, plus broader “world knowledge” and more natural conversations. Apple is pitching it as a much more capable assistant — one that could make AI feel normal for people who’ve never really used it before. But not everyone gets the same Siri. Some of the most advanced on-device features are limited to Apple’s newest hardware, including iPhone 17 Pro, and users have to join a waitlist even in beta. Tom asks the big question: if Siri finally works the way Apple has long promised, could that genuinely change how we use our phones — or is this still another catch-up move in the AI race? The Interface is your weekly guide to the tech rewiring your week and your world. Hosted by journalists Karen Hao, Thomas Germain and Nicky Woolf, each episode unpacks, week by week, how technology is shaping all our futures. No guests. No jargon. Just three sharp voices debating the stories that matter — whether they shook a government, broke the internet, or quietly tipped the balance of power. New episodes drop every Thursday on BBC Sounds in the UK. Outside the UK, find us on BBC.com or wherever you get your podcasts, or watch the video version on YouTube (search “The Interface podcast”). To get in touch with the team: theinterface@bbc.com The Interface is a BBC Studios production. Producer: Natalia Rodriguez Ford Executive Editor: Philip Sellars See more episodes Programme website

Commercial and fishing vessels are required under international law to be fitted with systems that transmit data, showing where they are. But many vessels simply do not broadcast their location. Effectively, going dark. For more than a decade the non-profit organisation Global Fishing Watch has been using technology to shine a light on issues such as illegal fishing, sanction breaking, and potential human rights abuses at sea. Reporting for BBC Tech Now, Grace Ekpu finds out how AI and satellite data are being used in these operations. This video is from Tech Now, the BBC's flagship technology programme. Katty Kay and economist Kathryn Anne Edwards on why some companies might use AI as an excuse for normal firings. Rising artist Shaqúelle Whyte on why painting for him is all about process, repetition, and doubt. Vance Howard of Howard Capital Management says he is optimistic after the Federal Reserve chair's debut. Florida's natural springs are portals to a prehistoric world hidden beneath turquoise and deep-blue waters. A local guide takes us to the layered slopes of Bokty mountain and on the multicoloured paths of Kyzylkup canyon. An unassuming German city had been scoring among the top three for the last two years. This 16th century concept originated in tea ceremonies and martial arts and is still engrained in Japanese life. What does a fish with lungs and a beetle larva have to do with survival? Leah Bennett of Concurrent says even with a US-Iran deal, it will take time for oil prices to normalise. Gus Scacco of Advisors Capital Management says Kevin Warsh will, for now, continue with the 'Powell scenario'. Millions of mummified animal gods show how ancient Egyptians turned to animal worship out of desperation. Chris Brigati at SWBC says even though volatility is not over yet, we are past the oil price peak. Danish adventurer Thor Pedersen spent nearly a decade reaching every country without boarding a plane. Katty Kay and Ryan Burge on why some Americans now say religion matters more to them, without going to service. The Nigerian-American soprano on how the stage allowed her to feel everything she was taught to hold back. Markets open higher as chip stocks bounce and oil price rises amid the US threatens to hit Iran 'very hard.' SoFi Stadium's 360-degree double-sided infinity screen with immersive tech is redefining the fan experience. Julian Koski of New Age Alpha says he is not worried about rising inflation as 'these events don't last long'. This Los Angeles venue uses cutting-edge technology to bring the World Cup experience to life. Mackenzie Peterson at AdvisorShares says the OpenAI IPO shows continued 'hunger and desire' around AI.

Wearable AI can help travellers navigate cities, translate menus and fundamentally transform travel. But a weekend in Paris showed me the trade-offs behind the convenience. In Paris, I find myself standing beneath the Eiffel Tower having an argument with my glasses about how tall it is. The first time I ask, the answer comes back as 330m. A few minutes later, I try again. This time, the glasses confidently tell me, through a tiny speaker near my ear, that it is 324m tall. This is not the kind of discussion I had expected to be having during weekend testing Ray-Ban Meta AI glasses as a travel companion. This small discrepancy in height – the official Eiffel Tower website lists it as 1,083ft, or 330m – throws up some big questions for me. If I can't trust the glasses to tell me the correct answer to something easily verifiable, what can I trust them with? Launched in late 2023 by Meta and EssilorLuxottica, the parent company of Ray-Ban, the glasses are part of a fast-growing category of AI wearables that is moving from novelty to mainstream use. More than seven million Meta smart glasses were reportedly sold in 2025, and rival products from companies like Google and Samsung are now in development. Apple is also widely reported to be working on smart glasses of its own. The promise for travellers is immediately appealing: live translation, help with directions, hands-free photography and quick answers about restaurants, landmarks, menus and whatever else happens to be in front of you. In theory, they combine the functions of a guidebook, translation app, smartphone and audio guide in a single wearable device. As someone who spends too much time in new places navigating with Google Maps and staring at her phone, I am excited to see whether they make me feel a little less like an awkward lost tourist as I wander around Paris. But I am also uneasy. Camera-equipped smart glasses have attracted growing criticism for the way they can be used to film and photograph people without their consent. I wonder whether this particular type of travel tech will help me connect more deeply with the city, or mark me out for all the wrong reasons. The Ray-Ban Meta AI glasses I packed for my weekend trip to Paris are chunky and black-framed. They look like a standard pair of glasses but I find speaking to thin air a little strange, so I hide my mouth behind my hand, as if I am about to cough, before asking a question. Unlike augmented-reality headsets, the glasses do not project information across the lenses. Instead, they use cameras, microphones, speakers and voice AI to create a kind of audio-first layer between you and the world around you. Wearers can make phone calls, take photos, record videos, listen to music and podcasts, translate text and speech, ask questions about what they are seeing, set reminders and request recommendations. I begin by asking the glasses to give me directions as I walk with my travel companion from the Gare du Nord through Le Marais and down to the Seine. They quickly send the route to the maps function on my phone and alert me to open it. I suppose it's technically what I asked for, but it means I have my head down, staring at my phone, instead of looking around and enjoying the city, which is not what I hoped for. That said, the experience improves dramatically when realise I can set a route on Google Maps and follow voice directions through the glasses, using them as a headset rather than a navigation tool. That works surprisingly well, and it's better than wearing headphones because I can also hear everything around me. At Place de la Concorde, I ask for a brief history of the square. The glasses tell me that Louis XVI and Marie Antoinette were executed here when it was called Place de la Révolution. So far, so good. Later, my travel companion asks me how long it took to build the Louvre. I tell her, via the power of the glasses, that it took more than 800 years. This becomes one of the device's most useful functions, acting like travel guide for my ears, able to pull up information quickly and efficiently without forcing me to scroll through my phone or flick through a guidebook. Instead of looking like a lost tourist, I can keep walking while satisfying my curiosity. I try out some other applications on the go. I ask it to convert currencies to give me a ballpark idea of prices, and it is successful. The translation function works well too, when I look at a French newspaper and ask for an English translation. Later in a café, the glasses provide both a verbal summary of the menu and a line-by-line translation. As someone who speaks French, I find myself nodding along. But at the Eiffel Tower, things start to unravel. Suddenly I'm questioning everything that has happened so far today. The issue isn't the size of the mistake so much as the confidence with which it was delivered. When I ask where the information came from, the glasses respond vaguely: "I get my information from my training data, internet searches and other sources." This leaves me wondering how I am supposed to judge which answers are reliable. All sources seem to be treated as equally valid, whether they are accurate or not – a known AI issue in other areas. Still, it handles most of my basic travel-guide-style questions well, and when I later verify the answers, it is correct in all other cases. The technology is useful – provided I have time to check the information that matters. Some features are less convincing. While I can take photos and videos from my point of view, I can't zoom or focus manually, and unless the glasses sit perfectly level on my face, the results can be slightly crooked. I far prefer the control and options available on a smartphone camera. The object-recognition feature is similarly mixed. Sometimes it identifies what I am looking at accurately, while other times it just offers vague observations such as "you're looking at a view of a street, maybe in Paris, with tall buildings". Which is true, but not exactly revelatory. And then there are the privacy issues. Wired has reported instances of Ray-Ban Meta AI glasses being used to record women without their knowledge or consent, earning the glasses the ugly nickname "pervert glasses". Swedish newspaper Svenska Dagbladet also reported that workers reviewing data from Meta smart glasses had seen intimate and sensitive footage, including people who appeared not to know they were being recorded. More like this: • Smart glasses are 'an invasion of privacy' - Meta's are selling better than ever • The best Christmas market in Europe • David Lebovitz's ultimate guide to the best bakeries in Paris right now Meta's own privacy policies state that voice interactions can be stored and processed using machine learning and trained reviewers to improve its products. Its voice privacy notice also says transcripts and stored audio recordings may be kept for up to a year unless users delete them sooner. Those concerns are serious. But as a traveller, my biggest reservation turns out to be something else.  The longer I wear the glasses, the more I feel they undermine one of the things I value most about travel. In years gone by, if I was lost, I might have asked a stranger for directions or relied on a hotel concierge for recommendations. I might have wandered aimlessly in the city and discovered something through serendipity. With the glasses, I don't need to ask anyone anything, because the glasses handle it all. The technology deals with directions, translation, information and recommendations. It reduces issues and increases convenience. But the glasses don't live up to the promise of helping me connect well to the world around me. If anything, they sometimes place another invisible layer between me and the city. This feels like part of a wider issue in travel technology, where every innovation promises to make travel easier. Yet some of the most memorable moments on a trip happen precisely because things don't go according to plan. It has me thinking about what I really want. Swapping accidental discoveries and human interaction for convenience feels like too much of a trade-off. I can see the appeal for trips where language, logistics or cultural context are more challenging – for example, an upcoming trip to Japan, where translation and basic travel-guide functions could genuinely improve the experience. But for future trips around Europe, where I already have some familiarity with the language and customs, I'm less convinced they would add enough value for me. They might also make sense for business travel, when time is limited and convenience matters. That said, I would still verify important information, and I remain uneasy about what Meta is doing with my data. My advice is to stay conscious of how and why you're using them, and what you want to get out of your trip in the first place. As wearable AI becomes increasingly woven into daily life, travellers will need to think carefully about what they want technology to do for them. The glasses excel at making some elements of travel easier. The question is whether easier always means better. -- If you liked this story, sign up for The Essential List newsletter – a handpicked selection of features, videos and can't-miss news, delivered to your inbox twice a week. For more Travel stories from the BBC, follow us on Facebook and Instagram .

"Everyone's really upset mum - loads of them have got their own YouTube channels." That was my 12-year-old son's summary of how the news about the social media ban for UK under-16s went down in his classroom. Exactly how a bunch of 12-year olds might have ended up with their own channels in the first place when the minimum age is supposed to be 13 shows just how big a change in culture the government is trying to make. In Preston, school pupil Isabella went viral when a BBC colleague asked her on-camera what she would do instead with the nine hours of screentime she had racked up over the previous weekend: "stare at the wall," she deadpanned. The exact logistics of the ban have yet to be set out but it is very possible that its introduction will herald the biggest ever change in the UK in terms of how everyone, children and adults alike, accesses the internet. Millions of us might have to share some official ID which includes our date of birth, in order to access a whole range of platforms from next spring. The ban has been broadly welcomed by campaigners, including a group of bereaved parents who say their children died as a result of a variety of harms on social media. But for others, what the government is planning goes beyond getting the nation's kids to spend more time off screens and engaged in alternative pursuits (even if that does include staring at walls) and amounts to a profound reshaping of how it is assumed young people will accumulate fresh knowledge and also how the rest of us will move around online. There is the potential impact on education. "YouTube is where we all go to learn," says Dr Tom Crawford, aka Tom Rocks Maths, who shares maths skills with his 250,000 subscribers on YouTube, which is included in the ban. "And that includes teenagers." So, are we really witnessing the profound shift that some claim? And if we are, how will it reshape our relationship with the online world? Much of the concerns raised so far about the proposals have been about civil liberties and government overreach. But there are other, more prosaic, unintended consequences to consider too. "Every young person I have spoken to has told me the same thing: they will find a way around it," says Paddy Crump, campaigns director at Flippgen, a youth-led non-profit group that goes into schools to try and help young people build healthier relationships with the online world. That is certainly what seems to have happened in Australia, where seven out of 10 children aged under 16 who had a social media account before it introduced its ban in December 2025, still have some access, according to a report by the country's e-safety commission. Crump argues that the measures offer "false hope dressed up as protection" and will simply shift young people's online behaviour elsewhere: including to smaller digital platforms which fly beneath the radar of regulatory scrutiny. "There are some pretty dangerous places for children and teens that make Instagram look like Disneyland," notes Ari Lightman, professor of digital media and marketing at Carnegie Mellon University. And critics of the proposals warn of other unintended side-effects. Crump fears the ban could make young people less likely to seek support for online harms if they do encounter them, as well as isolating them from communities and information. One teenager sent me a message to say that without social media they would not still be here: the friendships they had made online had given them reasons to continue living. Some parents with SEND children say social media and watching videos is their primary way of engaging with the world. An online e-petition is calling on the government not to ban social media for under 16s "because for many young people social media is how they communicate with their friends. Some people view social media as a lifeline". It has gained more than 100,000 signatories in the past few days. Home education message boards are also lighting up with parents concerned about how to navigate the ban while teaching their children away from schools. "I learned to tie a bow tie by watching a tutorial on YouTube," says Crawford. "What if you're an 11-year old that needs to wear a tie to school for the first time? What if you want to know how to apply makeup and there's no-one at home to show you? What if you're worried about your upcoming GCSE exams and want to check how to answer a question on bearings? This is what a ban on YouTube takes away - the ability to learn." Older generations might retort that they managed to acquire all this knowledge without the help of the internet. But that ignores how fundamentally teenagers have become accustomed to using not just YouTube but also other social media platforms as a research tool. SEO expert Mehwish Malik from Link Builder says the younger end of Gen Z (aged 14-29) use TikTok as a search engine: their preferred gateway to information and to trusted brands. So how can all this be addressed? The government says this is for the tech companies to figure out. "If YouTube wants to come up with something that's an intermediate option that allows that young person who wants to watch history documentaries to watch them but isn't then getting all of these short reels, that's a different proposition," said education secretary Bridget Phillipson on the BBC's Newscast. Industry sources argue that technically it's not that simple to set something like this up. "Ask the government!" messaged one when I posed the question about how it might work. Parents could of course just choose to sit down and watch something with their child using their own accounts if they have the time and willingness: YouTube claims that half of UK users watch its videos on the TV at home, with multiple sign-ins available. "As I see it, the main issue here is that YouTube isn't social media," says Crawford. "YouTube is the 2026 version of television." With the design features that aim to keep people on the platforms for as long as possible also under review for additional measures impacting 16 and 17 year olds, perhaps social media will end up withering on the vine because it just won't be interesting enough for young people to engage with even when they do reach the right age. "If you are drinking a glass of wine and it magically keeps refilling without you noticing, you will just keep drinking. Your brain only 'wakes up' when you reach the bottom of the glass," says Asa Raskin, who invented the concept of infinite scrolling 20 years ago. He now works at the Center for Humane Technology, which he co-founded, and accuses the tech companies of "weaponising" his idea. He says he intended to create "a seamless user experience" before the era of social media, and regrets that his invention has ended up being used "not to help people but to keep them hooked". The absence of young people could also change the social media experience for everybody else. MrBeast is arguably the world's most successful YouTuber with half a billion subscribing to his mix of challenges, stunts and charity. He started out at 13 and as a child studied the algorithm. He went on to corner the market in "watch time", created a factory of content and is now a billionaire. Would he have had the same idea years later? Professor Amy Orben is a psychologist at Cambridge University who has advised the government on screen time for children. She accepts that any ban will be "imperfect" but also agrees the government cannot do nothing; despite the evidence on social media harms itself being complex. There are of course acute and tragic cases, but broadly, she says the evidence for large populations links social media use to only a small decrease in mental health. In her opinion, the tech firms could help both regulators and themselves by sharing more about what they know from the billions of young people they see on their platforms day in, day out. "Social media companies have offered exceptionally little data on their internal research," she says. When it comes to age verification, it is expected that the tech giants will do the checking. "The methods available to platforms are well established. Identity document scanning with a face match, email-based age checks and facial age estimation are proven to work at scale," says Andy Lulham, Chief Operating Officer at Verifymy. That is a concern for those who worry about the reach of Big Tech into our lives - and that affects all of us, not just the young people who need to prove their age. Some see this as a major attempt by the authorities to control who can access what on the internet: this troubles privacy and rights campaigners as much as it relieves parents who are worried sick about what their children are being exposed to. For those in favour, this is a price worth paying to protect children. For Elon Musk, the controversial owner of X, it has a more sinister undertone: "The real goal is to enable the UK government to track everyone," he posted. It's not the first time the US trillionaire has waded into UK politics and he isn't universally welcomed when he does. Needless to say the government denies this. Musk is not alone in his concerns: an international campaign called Stop Killing the Internet also launched this week. The group, which includes the Index on Censorship and Big Brother Watch, is concerned that various forms of surveillance, as it considers this to be, limit rights to freedom of expression for children and adults. Silke Carlo, Director of Big brother watch said: "We want all children to be safe online, but these policies create new safety and privacy risks for young people and entire adult populations alike. Far from reigning in Big Tech in, age-gating policies gift corporations masses more of our personal information whilst letting them off the hook for their design choices". For Carlo, those risks include the potential for sensitive children's data, such as proof of age and face scans, to be stolen and misused. And then of course there's the potential for future mission creep. "'Keep children safe' can end, three statutory instruments later, as a duty to scan every message or verify every face, administered by a regulator the public cannot easily call to account," warns computer scientist Professor Alan Woodward from Surrey University. "A walled garden is only a refuge if the people inside chose the wall, can see over it, and may leave when they wish." I suspect my own 12-year-old son and his peers will spend a lot of time looking for potential exits from the walled garden they are about to find themselves in, even if it's supposed to be for their own protection. If the ban does come into force in 2027 as planned and they can not escape it, today's under 16s are unlikely to spend the following years staring at the wall (I hope). Child-free digital spaces will feel different for adults too: I think we might be on the cusp of a new social media era, one that is less intense. It might leave us all with more time to read books, go outdoors... or use our phones to chat with AI instead. Additional reporting: Philippa Wain BBC InDepth is the home on the website and app for the best analysis, with fresh perspectives that challenge assumptions and deep reporting on the biggest issues of the day. Emma Barnett and John Simpson bring their pick of the most thought-provoking deep reads and analysis, every Saturday. Sign up for the newsletter here

The company behind the UK's newest video game console is not concerned with the latest state-of-the art graphics. Instead David Lee, the chief executive of American technology firm Nex, tells me its cube-shaped machine, the Nex Playground, is all about getting kids active. Launching in the UK and Ireland on 22nd June, the Playground ditches controllers for body movements, tracking players using AI and a built-in camera. The relatively little-known device surprised the games industry when research firm Circana revealed it was the third best-selling console in the US over Black Friday 2025, outselling the Xbox Series S and X. While motion-tracking in gaming is nothing new - the Nintendo Wii came out in 2006 - how long children spend on screen time is still a hot topic among many parents and governments today. Ahead of the UK launch I spoke to parents who already own the console in the US, and tried the machine myself to find out how it works - and if it really could get families feeling fitter. When it is released in the UK and Ireland the Playground will cost £269 (€319). While users get five starter games to try out for free, a subscription is needed to access most of the Playground's 60-plus games, which include tie-ins with kid-friendly favourites such as Peppa Pig. A yearly game subscription is £90, while a quarterly one is £45. Nick from Louisiana, who has had the Playground for six months for his children aged three and five, said the subscription was his "biggest hesitancy" when he first began researching the device. "But when you consider the fact that a single Switch game costs about $70 or $80, it's really not too egregious," he said. Brian, a parent from Philadelphia who bought the device a month ago for his six-year-old son, agrees. "I do think there's plenty of value here, especially when you consider the dollars per hour of this activity versus many others," he said. "The initial set-up was extremely smooth and the interface is simple and easy," said Corey, a parent from North Carolina who bought the device a month ago for his children, aged seven and four. The system uses AI and its wide-angle camera to track 18 points on the player's body to create an on-screen matching avatar. According to Brian, the camera tracking technology sometimes felt "a little lacking" and less precise than older systems like the Wii or Xbox Kinect. The camera quickly configured itself to fit around me and my not-so-large living room area, so I could slice Fruit Ninja's flying produce with my bare hands and hit (most of) the notes to the sound of A-ha's Take On Me in the rhythm game Starri. While some games felt "like tech demos" according to Corey, others felt more substantial. "The subscription lets me not worry about any of that and just dive into whatever my family wants to explore," he said. As the console effectively puts a camera in people's living rooms, Lee said player privacy was the "number one priority" for his company. "The camera is only for tracking motion; we don't save the video anywhere; it is processed in real time, locally on the device, rather than in the cloud" he said. The Playground has a kidSAFE+ COPPA certification, something which ensures it complies with the Children's Online Privacy Protection Act (COPPA) , a US law regulating how companies collect and use personal data from children under 13. The camera also comes with a lens cover, and parents can hide games or music they think their children are not old enough for yet. The Playground presents itself as "an alternative to passive or open-ended digital entertainment" amid "a growing national conversation around children's screen time habits". Anyone using the device will still be looking at a television screen, so the benefits for children may be more a "compromise" to include some healthy activity alongside it, Nick said. The parents I spoke to said their children often played on the console for between half an hour to an hour in one session, with the games typically being used as a way to transition into another activity, or to allow some structured play. Brian said while getting a Playground meant they "compromised on increased screen time", the games were still engaging in a way that he believed "typical cartoons or movies" were not. So did I work up a sweat in any of the games I played? In the initial starter pack, three games involved only a small amount of moving my arms. The final two, the rhythm game and a set of mini-games, did include more full body movements. The fuller Play Pass has a specific "Health & Fitness" category which includes sessions such as daily Zumba workouts, complete with an on-screen instructor shouting out movements in time to the music. It wasn't clear if I was managing to hit every motion correctly, but it did at least get me moving a bit more, and felt more convenient than heading to a gym class. While the Nex Playground may not be a direct contender to major consoles given its target audience, its recent performance during Black Friday showed it can nevertheless still hold its own in sales. According to Nex, the cube has now surpassed a million lifetime units sold since its launch in December 2023 in US and Canada. Chris Scullion, deputy editor of Video Games Chronicle, bought the cube seven months ago for his daughter. He said the device would probably never "realistically challenge" something like the Nintendo Switch 2, which also has family and children as a target audience and had sold over 17 million units by the end of 2025 . But he added the system's "clear family focus" could arguably make it a "more compelling offer" for parents looking for a modern Wii Sports or Wii Fit replacement. Alongside its UK and Ireland launch, Nex has also announced a multi-year partnership with Wrexham AFC that will bring Nex branding to the club's kit sleeves, fan activations at the Racecourse Ground, and community programming. If that strategy pays off, Nex Playground may find its biggest success not as a rival to consoles, but as part of a wider push into how families play, connect and spend time together. Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here .

The latest instalment in Rockstar's blockbuster game franchise, Grand Theft Auto, is set to be the biggest games launch of the year. Details are still scant, although we do now know that GTA 6 will be available to pre-order on 25 June, the developer has announced . Analysts believe Rockstar's action adventure could become the most expensive game ever made, with estimates putting development costs at more than $1bn (£866m). We're still awaiting some crucial information about the game - but here's what we do and don't know about GTA 6 so far. GTA 6 will be released on 19 November - and with Rockstar recently opening up pre-orders, this date seems all but certain. But it was delayed twice previously , from autumn 2025 and May 2026. In a statement when the second delay was announced, Rockstar said it needed extra months to finish the game with the level of polish fans had come to "expect and deserve". The sixth game in the main series will be released on PlayStation 5 and Xbox Series X and S. There's no news on when a PC version will be released - it's also currently unknown whether it will come to Nintendo Switch 2. Rockstar and its parent company Take-Two has still not revealed its price. It's something that will likely only be revealed when pre-sales begin in late June. Some analysts have estimated the game could be the first to be priced at $100 (£76). Rockstar has not yet confirmed whether an updated online mode will be released as soon as GTA 6 launches, although given the success of GTA Online it is likely to exist at some point. How any future age verification legislation in the UK around social media and particularly messaging in gaming could therefore affect it, if at all, is still unknown. GTA 6 will feature its first ever playable female protagonist in a 3D setting - Lucia - alongside her partner in life and crime, Jason, as a second playable character. From what we've seen so far in the two gameplay trailers , the story will follow the pair in a Bonnie and Clyde-esque adventure through the seedy underbelly of America after an " easy score goes wrong ". Like other GTA games before it, the sixth chapter is set in a fictional US state - this time in Leonida, which is Rockstar's version of Florida. Fans of previous games will be excited to hear that Vice City, which is inspired by Miami, will return, featuring as the main city within the state. GTA 6's cover art revealed by Rockstar on 18 June made several nods to the setting with alligators and flamingos. It's no secret that GTA 6 has been a few years in the making. The franchise's last instalment, GTA 5, was released in September 2013 - and quickly became one of the best-selling games of all time. It's unclear whether Rockstar has been working on the sixth instalment for all of that time, however, as in 2018 it also released another huge sequel - Red Dead Redemption 2. Rockstar confirmed for the first time that it was working on a sixth game in the series in February 2022. The rising cost of development and immense pressure to live up to the series' hype are all contributing factors as to why gamers have had to wait so long for the game's release. Several months later, footage and images leaked after teenage hackers targeted the company . It has also seen some issues since, including accusations of "union-busting" by sacked workers at Rockstar North , its Edinburgh HQ, where developers have been racing to get GTA 6 ready for release in November. To put it simply, the GTA franchise is one of the biggest and most profitable entertainment properties in history. GTA 5 has sold nearly 230 million copies, making it one of the best-selling games of all time - which has generated billions of dollars in revenue for Rockstar. The game's sandbox gameplay, where players can explore vast open worlds with considerable freedom - sometimes controversially - has seen it be continuously lauded as the ultimate expression for what the interactive medium can do. Freelance video games journalist Vic Hood said each entry in the series "continued to push technical and gameplay boundaries", with GTA Online helping to "pioneer the live service model as we know it". She added that the games were also "extremely culturally relevant" too. "Rockstar has always had its finger on the pulse of music, entertainment, societal, political, celebrity, and online culture trends, and they've never shied away from satirising these things in GTA games," she said. Speculation as to how a sixth game can top even the gigantic heights of its predecessor has meant that any latest news, no matter how small, has been met with a flurry of hype and excitement. Both of GTA 6's previous gameplay trailers currently have a total of roughly 447 million views combined. Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here .

At an altitude of almost 4,000m (13,000ft) and receiving almost no rainfall, the Himalayan village of Sakti is a hostile place to be a farmer. "Ladakh has a brutal, single-cultivation season," says Gelak Gutme, who has been growing wheat, peas and potatoes there for most of his 65 years. "It is a desert with an extreme climate," he says. Conditions have become worse in his lifetime. Global warming means that the smaller, low altitude glaciers they relied on to water their crops have disappeared. "Now there is scarcity of water. Last year I lost everything - my entire field got dried due of lack of water," Gutme says. "For generations, small glaciers sitting right above the valleys acted like frozen water towers, holding onto water all winter and releasing it right when spring farming began," explains Lobzang Fardod, who is a member of a local water management committee in Ladakh. "Now that those lower glaciers have completely vanished into a desert of dry rock, there is nothing left at the top to melt," he says. The mountain summer is short, so farmers have to plant their crops by May, otherwise the crops will not be ready before the winter hits again. A reliable source of water in early spring is crucial for them. To secure that vital resource, in the early 2010s some Ladakh villages attempted to create their own reservoirs of ice. The system involved piping water from higher up in the mountains during the winter and spraying it into the air, where it would freeze, and over time form large towers of ice, called ice stupas. They successfully supplied melt water in the spring, but were a "nightmare" to manage under harsh winter conditions, says Fardod. If temperatures dropped quickly below minus 20C, or sometimes minus 30C, the water in the pipes was liable to freeze, cracking the pipes and ruining the whole system. To guard against that, during the winter teams of four or five farmers would camp high-up, near the water source, rushing to any potential blockages with boiling water, often during the night when temperature drops were most likely. But enduring those freezing, winter nights high in the mountains could be phased out. "Because traditional water systems are failing, Leh-Ladakh has become a hub for innovative, grassroots hydraulic engineering," says Murtaza Ali, executive engineer in the Irrigation and Flood Control Division, at the Ladakh Autonomous Hill Development Council. Leh is the capital of Ladakh, a disputed region in Indian-administered Kashmir that is sandwiched between China to the east and Pakistan to the west. More on Ladakh's political challenges As well as the potential for cracked pipes, the ice stupa system was not very efficient, says Ali. Because water flowed constantly, on warmer days fresh water would melt the ice that had already formed. But over the last couple of years that method has undergone a tech upgrade. In partnership with private company Acres of Ice, a new system has been developed which precisely controls ice production. Called an Automated Ice Reservoir (AIR), the process also involves piping water down from higher up in the mountains. The water arrives at the valley floor under pressure and shoots out of a vertical nozzle like a "massive fountain", says Dr Suryanarayanan Balasubramanian, the founder of Acres of Ice. That flow is computer controlled from a weatherproof control box, powered by solar panels and a battery. The control system is connected to a weather station, which monitors, environmental conditions, including the water temperature inside the pipe. If the sensors detect that the air temperature is dropping too fast, or the water temperature inside the pipe approaches a dangerous threshold, the control system takes action. It shuts off the valve at the top of the stream and opens a valve at the bottom to completely drain the standing water out of the pipe. That avoids the ruinous problem of cracked pipes, but the system is also more efficient at creating ice. Instead of continuously spraying water, AIR fires a burst of mist, coating the existing ice, and then shuts off. "The system waits precisely long enough for that layer of water droplets to freeze solid based on current wind and humidity, then fires the spray again," explains Balasubramanian. He says that AIR converts almost all of the diverted water into ice. The whole system runs automatically and uses a local wireless network to connect the control box and the various valves. But the villagers do have a manual override, if needed. It all appears to be making a difference to village life. "When we speak to the villagers, they are saying the groundwater is getting recharged and spring sources are getting revived. They are getting water in time. We are also planning a scientific study now to see exactly what impact it has made," says Ali. During the winter of 2025, Acres of Ice and the local government ran 10 AIR projects across Ladakh. "Our biggest challenge right now is to push the envelope in the technology to see how we can multiply the number of ice reservoirs we are building. With the same system that previously used to build only one ice reservoir, can we build a dozen?," says Balasubramanian. Back in Sakti, farmer Gutme is more optimistic about the future. The single AIR system has created a more reliable water source and he hopes the village will build at least two more of the artificial glaciers. "I am a farmer, land is all that I have to survive on. I don't know the technology, all that I know today is that I have water to grow my crops. "We live in harsh climate that makes our life difficulty and lack of water was creating more issues. Many of youths in the village wated to go to cities to work. That would have been a disaster."

AI will lead to more need for workers rather than make people redundant, Amazon founder Jeff Bezos predicted during an appearance at a tech conference in Paris. Bezos pushed back against growing concerns that AI will replace large numbers of workers. Instead he argued that the tech will unlock new opportunities and increase demand for human labour. This is in contradiction to some other tech and political figures - including former UK prime minister Rishi Sunak, now an adviser to Microsoft and AI firm Anthropic, who recently said AI was having an impact on young people's job prospects . "I know there's a lot of concern that many people have, including many smart people, that AI is going to make humans redundant and so on," Bezos said. "I totally disagree with this point of view. And I think, in fact, AI is going to create a labour shortage." He painted an optimistic picture of AI's future role in society, suggesting that people are limited not by a lack of ambition, but by barriers that technology can help remove. Billionaire entrepreneur Bezos was speaking about his new AI venture Prometheus, which is focused on accelerating physical manufacturing - a sector which is becoming increasingly automated . The UK's Trades Union Congress has warned that AI technology could repeat "the disaster of deindustrialisation" as shareholders get richer while jobs are "degraded or displaced". But it adds that AI could have transformative potential if developed properly, and workers could benefit from its productivity gains. Bezos also used his appearance at Europe's largest tech expo VivaTech Paris to outline his long-term vision for space exploration. He described space as "supply constrained, not demand constrained", arguing that access to space remains the biggest obstacle to future development. The Moon, he said, offers a natural starting point for humanity's expansion beyond Earth because of its proximity and resources. "We're going to the Moon to stay, not just to visit," Bezos told the audience, adding that technologies such as electrolysis could eventually allow lunar resources to be used to refuel rockets and support a permanent presence beyond Earth. The discussion also turned to another Bezos venture, space travel company Blue Origin. It had a recent setback after an uncrewed New Glenn rocket exploded during a ground test at Cape Canaveral in Florida in May. "It was a gut punch for the whole team. But what we've learned since then is we got really lucky," Bezos said. No injuries were reported in the explosion, and Bezos noted several critical pieces of launch infrastructure survived the incident, including propellant and fuel systems that would have taken significantly longer to replace. On the same stage as Bezos, Blue Origin chief executive Dave Limp said reconstruction work at the launch site is already underway and the company expects launches to resume before the end of the year. Blue Origin is in the race to establish itself as a major player in commercial spaceflight and lunar exploration, competing with Elon Musk's SpaceX in the growing market for extraterrestrial infrastructure. Away from the main stage, Unitree's humanoid robot was the definite crowd-pleaser. Constant queues of visitors wanted to see the latest advances in the robotics field. This time the robot was teaming up with French neuro-AI company HABS, which showcased technology designed to allow humans to interact with machines using cognitive signals rather than speech. The robot responded to commands generated through brain activity, via a headband with an electroencephalogram (commonly known as an EEG) attached to it. The test uses small, metal probes called electrodes that touch the scalp. The demo offered a glimpse of how future humans and machines could work together in the future. It also reflected a broader trend running through this year's event: AI moving beyond chatbots and into the physical world. Humanoid robots are increasingly becoming a reality with companies racing to develop machines capable of working alongside humans in healthcare, manufacturing and hospitality. Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here .

More on this story.

Snapchat's parent company has announced it is releasing new smart glasses, a decade after its original pair lost the company tens of millions of dollars . The new augmented reality (AR) glasses, called Specs, will allow users to see digital elements overlaid onto the world. They will cost £1,995 in the UK and $2,195 in the US when shipping begins this autumn. That makes them cheaper than Apple's Vision Pro mixed-reality headset and its $3,499 starting price, but far more than Meta's smart glasses, which start at $224. Evan Spiegel, co-founder and chief executive of Snap Inc, said the glasses marked the "beginning of a new era in computing". Ben Hatton, a market analyst at FDM CCS Insight, said the price tag meant the technology was "unlikely to become a mainstream device any time soon". Snap's core audience of younger consumers "rarely have this sort of money to spend on a single gadget," he said. A $200 refundable deposit is needed to pre-order the glasses, which will ship in the US, the UK, and France. Snap says they are designed to be "wearable for everyday life," though the battery lasts four hours on average before needing charging again. A charging case holds up to 20 hours of battery life. Unlike Meta's Ray-Ban and Oakley models, the glasses function without needing a "tether" to a smartphone. And they differ from Apple Vision Pro - which is mainly designed for use at home - as they do not require a wire connected to a "puck", or battery pack. Hatton said while this improved "wearability and mobility" for the user, it came "at the cost of lower battery output". "Despite the impressive features and experiences available through Specs, glasses with a 4-hour mixed-use battery life and bulky design are not going to replace the smartphone any time soon," he said. Users will be able to use the glasses' AI assistant to complete tasks such as getting directions or asking questions about objects they are looking at. They will also be able to watch videos, browse the web, play AR games and record what they see through the glasses. Smart glasses have been criticised for a lack of privacy, having been used to film women in public without their consent . In February, UK data privacy watchdog the ICO wrote to Meta, after revelations data workers in Kenya had to watch videos of people having sex and using the toilet which were filmed with Meta's smart glasses. "Privacy has to be built in from the very beginning," Spiegel said, adding: "Specs only work if people trust them". A built-in light will glow when the device is recording, with Snap stating users will be given control over what data gets stored, synced, shared, or deleted. Snap Inc shares fell by some 9% after the announcement at the Augmented World Expo conference in California on Tuesday. This is not the first attempt at smart glasses for Snap. In 2016 it launched Spectacles , camera-equipped sunglasses which cost £129 and could record short video clips of up to 10 seconds. In 2017 the company revealed losses of $40m (£30m at the time) due to "excess inventory" of the glasses and cancelled orders. Updated versions of Spectacles were released in both 2018 and 2019 with a higher price tag (£330 for the third instalment) and upgrades including a better camera resolution. It has not released a widely available consumer product since then. Snap later began shifting towards AR, offering new Spectacles in 2021 to a "select group of global creators" in the field. In May, Google announced it would launch its own pair of smart glasses a decade on from the doomed Google Glass, which had been pulled after backlash on price and privacy concerns. Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here .

After the government's announcement on Monday, we know a social media ban is coming for under-16s in the UK . However, details on which apps are and are not included, besides those named by the government, and how the measures will extend to gaming sites like Roblox, remain sparse. And many are already asking whether enforcing the ban will mean cracking down on virtual private networks (VPNs), which can disguise someone's location online. Ministers have said they will provide an update on further restrictions like potential curfews, curbing of "addictive" features like infinite scroll and AI chatbots, in July. But here are some of the big unanswered questions about the UK social media ban. The government has only specified a handful of sites included in the UK's social media ban. These are Snapchat, TikTok, YouTube, Instagram, Facebook and X. However, it says its restrictions will closely follow Australia's - where the above apps plus Kick, Reddit, Threads and Twitch are banned for under-16s. So where does that leave Roblox and gaming sites like it? Roblox is hugely popular with children in the UK, but it has been accused of failing to keep them safe - including allegations it allowed children to be preyed upon by adults . Roblox says it has made changes to protect child users , including by expanding tools to estimate a user's age to determine the games and experiences they can access. "It is not yet clear how they will treat gaming sites," says Lorna Woods, professor of internet law at Essex University. "Though if they follow the Australian approach, these will lie outside the Australian ban." Regardless of whether it does end up being banned for under-16s, Roblox will still need to abide by requirements for all platforms to disable features that let strangers communicate with children, announced alongside the ban. The government made a point of saying this restriction would apply to gaming services. Children will still be able to participate in multiplayer online games, though, it adds . YouTube is going to be included in the ban, but YouTube Kids will not be. However it is not yet clear how YouTube would tackle children accessing the platform through search without an account once a ban is enforced or how educational content could be carved out from any restrictions. YouTube said it was a "vital resource for young people" and that a ban risked pushing children towards "anonymous, less safe services". But questions are emerging about what including YouTube in a ban would mean for children's access to educational videos on the site. Citing research carried out in partnership with consultancy firm Livity, Google says 95% of UK teens surveyed had said watching videos helps with school work. According to the government, there will be "a narrowly defined list of exemptions" to its definitions for platforms included in the social media ban to keep educational services available to children. One of the few platforms we do not expect to be included in a ban on social media for UK children is WhatsApp. The messaging app, owned by Facebook-parent Meta, is used by half of all 8-17 year olds, according to Ofcom. The government says in a press release about its plans for restrictions it is considering "how to determine exactly which platforms will fall under the restrictions". But it does not expect to include messaging apps like WhatsApp and Signal. According to the government, its ban will cover platforms "whose purpose is to enable social interaction and which allow users to post material". While this definition would likely not include messaging apps, it could, in theory, apply to Discord and Pinterest. Yet both, as well as platforms like Bluesky, Tumblr and Telegram, are absent from Australia's list of banned platforms. One of the biggest emerging questions about the UK's social media ban is how it will be enforced. The government says it will use "highly effective age assurance" - tougher checks that aim to accurately identify or estimate a person's age - to make a ban stick. These checks can include forms of verification like asking someone to scan their face (facial age estimation tech), upload a form of ID (photo ID matching) or use a digital identity service like Yoti. In a letter to the government , the regulator Ofcom says while it should be technically feasible, applying age assurance at 16 has its challenges. There are fewer methods available to accurately measure or estimate if someone is 16 or under than if they are 18 or younger, due to differences in ID or data available. For instance, Ofcom says credit card or email address checks can work at 18, but not at 16. "We have more work to do to understand the effectiveness and accessibility of different methods, the availability of identity and age attributes at 16, and the privacy considerations of different existing and emerging methods," it has told the government. These regulations were followed by a spike in downloads of VPNs and calls from some, including the Children's Commissioner, for age checks to restrict them for young people. VPN providers and privacy experts say doing so would mean having to collect user data to carry out checks, and negate privacy as the incentive for using them. Some now speculate tech-savvy children may use them to get around social media blocks. The government has not revealed any plans to regulate them, but ministers have said details about action alongside the social media ban, including regarding VPN use, will come in July. Children's minister Josh MacAlister told the BBC there were "options there about whether we could age-gate VPN use, which would be really welcome". The prime minister said it was inevitable some teens would try to circumvent a social media ban, but this did not mean it should not legislate to protect current and future generations of UK children. Stop children using VPNs to watch porn, ministers told You need only look to the years it took to pass the UK's existing set of duties and laws for online platforms, the Online Safety Act (OSA), to see that regulating the fast-moving tech sector can take time. But where the OSA is still being implemented, the government recently used the Children's Wellbeing and Schools Act to give ministers powers to introduce further restrictions as regulations. This means the government won't have to bring a full Act of parliament to implement a ban, so it can act relatively quickly. Tech Secretary Liz Kendall told MPs on Monday: "I want a vote on it by the end of the year, and I want it to come into force as early as possible in the first couple of months of 2027". However if tech companies decide to challenge the government's process in the courts, using judicial review, that could slow down the timeline for a ban. "Whilst primary legislation is effectively immune from challenge, secondary legislation is subject to review by the courts as it lacks the rigorous multi-stage scrutiny required for statutes," says Giulia Carloni, senior associate at law firm Winston Taylor. Additional reporting by Chris Vallance Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here .

A social media ban for under-16s has been announced by the UK government. It will be introduced in early 2027. The government is also considering an overnight curfew and measures to stop infinite scrolling for under-18s. Keeping children off social media is the best way to keep them safe online, Prime Minister Sir Keir Starmer said. The ban will cover platforms including Snapchat, TikTok, Instagram, Facebook and X, the government said. YouTube will also be banned for under-16s, but YouTube Kids - a version of the video-sharing app designed to be safe for children - will not be included. A complete list of affected platforms has not been released, but the government said it would cover those "whose purpose is to enable social interaction and which allow users to post material". The government said it would "go further" than other countries. It means features including livestreaming and strangers being able to contact children will be restricted across a wider range of online services, including gaming sites. It means that although Roblox, the biggest gaming site in the UK for eight to 12-year-olds, will not be banned, certain features such as its chat function will be restricted. The government said it did not intend to include messaging services like WhatsApp and Signal in the social media ban. Most social media platforms already require children to be over 13 to create an account and use their services. The government said that for the "high risk" functions of livestreaming and chat with strangers, restrictions would be on by default for under-17s "to prevent a cliff-edge at 16". It said it was also looking into overnight curfews and breaks in infinite scrolling for under-18s, with more detail on this to be set out in July. AI "romantic companion" chatbots, which are designed to simulate sexual relationships or role play with users, will have to enforce a minimum age of 18. AI chatbots more widely will have to restrict "intimate functionalities" for under-18s, the government said. The prime minister said the government plans to pass regulations before Christmas. This would allow the ban to be introduced by spring 2027. The announcement follows a public consultation which received more than 116,000 responses. The government said "highly effective age assurance" measures would be used to check the age of those using social media. This typically involves requiring companies to use tech that accurately estimates or verifies someone's age - such as face scans or asking for ID. The regulator Ofcom has been asked to carry out a rapid study to identify the best ways to verify if someone is over 16. A number of platforms, including porn sites, are already required to carry out these checks. Ofcom has fined several platforms for not complying. But concerns have been raised about tools such as virtual private networks (VPNs) being used to get around these measures. Digital rights campaigners have also raised privacy concerns about requiring individuals to share their data with platforms. Technology Secretary Liz Kendall told BBC News the government would provide further detail on enforcement in July. Among those who have called for social media to be banned for under-16s are bereaved parents. Mariano Janin, whose daughter Mia died aged 14 after she was a victim of cyber-bullying, said the announcement had left him "speechless". "I think it's a change in the right direction, it won't be easy, but it will be possible," he said. Lisa Kenevan, who believes her son Issac died aged 13 after taking part in a viral trend, told BBC Breakfast that while parents wanted the ban "a lot sooner", the announcement means "we're in a good place now". However, speaking ahead of the announcement, Ian Russell - whose daughter Molly took her own life at the age of 14 after viewing harmful content online - accused the government of rushing restrictions for political reasons. He said "sledgehammer techniques like bans" only cause more problems. Some teenagers have expressed concern about how a social media ban will affect their ability to connect with friends and express themselves. George, who is 14, told BBC Radio 4's Today programme he felt people his age should be able to make their own choices because "we're beyond people saying what we have to do". Lilly, 15, from Cumbria, said apps like TikTok allow her to "communicate more", for example by allowing her to post videos of her performing. YouTube and Snapchat, and Facebook and Instagram-owner Meta, said a ban would not achieve the goal of keeping teenagers safe online. Instead, they warned, it could drive them to "less safe" platforms. Meta said restrictions should be "underpinned by an age verification system on devices", so that people are not asked to "hand over ID to dozens of individual services". YouTube said it was a "vital resource for young people". Snapchat said the majority of time on its platform was in private messaging between friends and family. Australia introduced a social media ban for children in December 2025. The UK government said it would follow this model. The most popular sites in Australia are now required to prevent children from being able to set up new accounts. Existing profiles belonging to under-16s had to be deactivated. Australia's rules apply to 10 platforms: Facebook, Instagram, Snapchat, Threads, TikTok, X, YouTube, Reddit and streaming platforms Kick and Twitch. The fact that it did not include gaming platforms Discord and Roblox was criticised by some. More than six months on, implementing the policy remains difficult. One student told BBC News that in her grade of more than 170 students, she knew of only three who had been booted off social media. And 70% of parents have told Australia's internet regulator their children were still on these platforms. No fines have been handed out yet but the eSafety Commission is investigating potential non-compliance by Facebook, Instagram, Snapchat, TikTok and YouTube. According to the Organisation for Economic Co-operation and Development, 25 countries have social media age restrictions in force, enacted, or under active consideration. In Spain and Portugal politicians have announced plans and partly voted the rules in. Their bans are likely to come into force this year. Governments in France, Malaysia, Denmark, Indonesia, Norway and Canada are also in the process of bringing in a ban. Laws have been passed in several US states but these are being challenged in the courts. Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here .

One of the most valuable assets owned by Lancaster University is stored in beer kegs. But it's not in one of the student bars. In a carefully locked laboratory rows of metal kegs are arranged on shelves and linked together with spindly copper pipework. The containers aren't loaded with prize beer but rather a gas called helium-3, one of the most expensive materials in the world. A single litre costs roughly $2,000 (£1,500), though the price can fluctuate. "The lab has been going for 50 years or so. Back then, the helium was quite cheap," says Dima Zmeev, senior lecturer. "Our very wise predecessors stocked up." In the near future, more people could be looking to build up such a stockpile. Helium-3 has applications in quantum computing and nuclear fusion. However, the main source of it today is tightly controlled – it comes from nuclear weapons. Specifically, from the decay of tritium, a form of hydrogen, inside those weapons. Around the world, tens of thousands of litres of helium-3 are likely to be produced this way every year, estimates David McCollum, distinguished scientist at Oak Ridge National Laboratory in Tennessee. But future demand could far exceed that supply. Some entrepreneurs and researchers say we need new sources of helium-3. It exists in the ground, though generally at very low concentrations. However, samples of moon dust, or regolith, from the Apollo missions suggest it may be present there at relatively high concentrations. As such, plans are now afoot to recover helium-3 from the moon. Helium-3 is an isotope of helium, defined by the number of neutrons in the atom's nucleus. Helium-4, with one additional neutron, is the comparatively cheap version – a gas that fills children's party balloons. Zmeev uses helium-3 in physics experiments. For example, he fills tiny chambers with the stuff, in a project to detect a type of mysterious dark matter particle. Should such a particle knock into one of the helium-3 atoms, it would make them all jiggle. This generates heat and that slight temperature rise can be measured. The helium-3 can be re-used again and again. Scientists mix helium-3 and helium-4 together at very low temperatures to create the lowest temperatures in the known universe, down to the millikelvin range (-273C). When helium-3 atoms gradually separate from a dilute mixture containing the two isotopes, they form a pure helium-3 layer on top. This separation is a phase change that consumes energy, inducing a cooling effect , like when steam evaporates off a cup of hot water. Helium-3-based cooling, or dilution refrigeration, is crucial for quantum computers. And helium-3 could also be used in some nuclear fusion reactors to one day create vast amounts of clean energy. One company planning to extract helium-3 from the moon is Interlune, based in Seattle. "We've spent the last four years developing, prototyping and testing technologies… We have a team of 30 people, and growing," says Rob Meyerson, co-founder and chief executive. Meyerson was president of Blue Origin, Jeff Bezos' rocket company between 2003 and 2018. One of Interlune's co-founders is Harrison "Jack" Schmitt, now in his 90s, who walked on the moon during the Apollo 17 mission. He has long advocated recovering helium-3 from lunar regolith. Interlune has tested some of its equipment during parabolic flights, in which a plane flies in a big arc to simulate zero gravity. The firm's kit could be integrated into a lunar lander as early as autumn 2027, says Meyerson. Eventually, Interlune aims to place autonomous, regolith-shovoelling excavators on the moon to scoop up the powdery material and process it. The idea is to crush and churn the regolith, releasing helium-3 contained within it. No-one knows with certainty what kind of helium-3 concentrations are present on the moon. Paul Burke, at Johns Hopkins Applied Physics Laboratory, says Apollo regolith samples might have lost some of their helium-3 on their return to Earth, skewing our understanding of how much is there. Plus, there might not be as many helium-3 hotpots as hoped, or they could be depths that are difficult to access. "It's important that we understand where the helium-3 is," says Burke. As Space News reported last year , lunar concentrations – perhaps between a few parts per billion (ppb) and 20-something ppb – could require excavating and processing hundreds of thousands of tonnes of the regolith just to obtain one kilogram of Helium-3. A "mountain-moving" prospect, says Burke. "We're not ignoring the fact that we've got to process large amounts of regolith," says Meyerson. Is the plan economically sensible? "We have run the numbers… for everything we need to get to the moon, extract the [Helium-3] and bring it back to Earth." Interlune declined to share those numbers with the BBC, or estimates for the total cost of developing its technology. Listen: Should we mine the moon? Another US company, Astrotech Corporation, has also announced its intention to go to the moon. In its case, via a SpaceX Starship rocket. Astrotech would extract helium-3 from regolith by heating it up. Tom Pickens, chief executive and chief technology officer says, "All of it is challenging." In previous space-based applications, his company made mass spectrometers, instruments that identify materials such as chemical elements and measure their concentrations. Work continues on a prototype for lunar helium-3 extraction and Pickens is bullish: "You'll see it." The company has "seven or eight" people working on the project, he adds. Quantum computers could eventually require thousands of litres of helium-3, depending on their design, suggests McCollum. He and colleagues recently published a paper scrutinising the energy and resource requirements of these devices. This means that the lunar helium-3 projects are already attracting interest. Helsinki-based quantum computing company has signed a $300m (£223m) deal with Interlune, for 10,000 litres of helium-3 annually from 2028-37. But there are alternatives. Some scientists are working on methods of cooling quantum computers that don't rely so heavily on helium-3, for example, points out Richard Easther at the University of Auckland. And helium-3 hunters might be able to recover useful volumes of from the Earth's crust after all. Pulsar Helium, headquartered in Portugal, is investigating the presence of helium-3 at a site in Minnesota. Concentrations there are around 12ppb says Peter Barry, a geochemist at Woods Hole Oceanographic Institution and a scientific advisor to the company. Conventional drilling could potentially yield helium-3 from the ground there he says, adding, "Minnesota is a lot easier to get to than the moon."

South Africa's fleet management market is projected to reach 3.8 million units by 2027, growing at 12.2% annually. Nigeria's logistics sector is expanding at a comparable pace. Yet beneath these headline numbers, fleet operators across Sub-Saharan Africa are hemorrhaging revenue through fuel theft, driver misconduct, vehicle misuse, and an almost total lack of real-time operational visibility. Fleet managers in Tanzania, Nigeria, and South Africa consistently report the same five pain points: they cannot see where their vehicles are at all times, they cannot verify who is driving, they cannot trust their fuel consumption data, they cannot coordinate across multiple depot sites, and they have no actionable analytics from the data they do collect. The result is that companies managing billion-shilling and billion-naira asset fleets are doing so with spreadsheets and phone calls — tools built for a different century. Phobolytics Technologies has engineered an AI-powered Fleet Intelligence Platform that eliminates every one of these pain points simultaneously. In-cabin computer vision cameras detect drowsiness, phone usage, seatbelt violations, and unauthorized drivers the moment they occur — not in a weekly report. AI-driven ANPR automatically logs and verifies every vehicle movement at every depot gate. Fuel anomaly detection cross-references GPS mileage, engine runtime, and consumption data in real time to surface theft or mechanical inefficiency instantly. Fleet operators already using AI intelligence platforms report 15 to 30 percent reductions in fuel costs and up to 40 percent fewer vehicle misuse incidents within the first 90 days of deployment. For logistics companies, construction firms, and government fleet operators across East and West Africa, the competitive gap between AI-enabled and traditional fleet operations will become irreversible by 2027. The operators who act in 2026 set the benchmark. Everyone else chases it. Phobolytics is actively partnering with fleet operators across South Africa, Nigeria, Tanzania, and Kenya. Contact the team at phobolytics.com to schedule a platform demonstration.

World Service , · 16 Jun 2026 , · 26 mins Available for over a year A group of cybersecurity researchers found a prompt which gets past ChatGPT’s guardrails and causes it to generate some disturbing images. We unpack what this tells us about the way AI is trained, and how it could be exploited. Also on the show, after a recent episode about potholes, we were contacted by the UK’s ministry of transport. We speak to their chief scientific adviser about potholes and the future of transport. And what is a quantum diamond magnetometer? We speak to the company which has just put one into space – in order to measure where magnetic north really is. Presenter: Chris Vallance Producer: Imran Rahman-Jones (Image: A phone with the white and black ChatGPT logo on it. In the background is green Matrix-style code. Credit: Getty Images)

Shiona McCallum visits Fifa HQ in Zurich to see some of the new tech being used in the World Cup. Martin Sharkey explores how changes in the way we watch content are shaping new technologies. More

Social media companies and authorities are failing women by focusing on nudity rather than consent when dealing with image-based abuse, according to a new report by gender justice organisation Chayn. Its criticisms are backed by Pakistani actress Ayesha Omar, whose experiences, along with those of other women, are described in the findings. One of the women at the centre of the report, whose name has been changed to Mahnoor to protect her privacy, explains that the images that changed her life were not nude. They were not sexually explicit. They showed a woman exposing her bare shoulders and wearing Western clothing. The 32-year-old from Pakistan told BBC Global Women that she returned to her childhood home when her marriage broke down. She hoped for comfort and support from her family, but instead, she and her young daughter were met with iciness. It's been over a year and her father and brothers still have not spoken to her. Colleagues at work who she has known for years will not look her in the eye. Mahnoor had expected a difficult divorce. It had never been an easy marriage. She says her husband, to whom she was married in an arranged match, was both verbally and physically abusive throughout their relationship. But it was the exposure of her private world that cost her the most. Like many young women, Mahnoor had saved lots of pictures of herself on her phone. She had taken photos of her everyday life - a nice dinner, a selfie when the lighting was particularly flattering. Many were years old. One was of her smiling after a new haircut. Another showed her on an overseas exchange programme with friends. Others were ordinary selfies, lying in bed, wearing a vest, with her eyes closed to show off her eyeliner. None had ever been shared publicly. She rarely posted photos on social media, mindful of the conservative culture of her community in Pakistan. According to Mahnoor, who is a university lecturer, her former husband gained access to her WhatsApp account and private images before distributing them to male relatives, colleagues and acquaintances. Mahnoor says he also cropped images of her with a group of friends, to make it appear that she was standing with a single man, insinuating that they were having an affair. The photographs, she says, were used to portray her as "a woman of bad character", an accusation that, in many communities, can carry life-altering and sometimes fatal consequences. With her friends and family, as well as colleagues, barely engaging with her, Mahnoor says she has lost her social standing and the once powerful position she held in her community. "I lost my voice," she told the BBC. "I no longer felt visible. "My family once respected me, my brothers respected me. Having your voice respected by your parents is such a great thing," she says. They used to ask for her advice, but that is no longer the case. Mahnoor's ex-husband has now remarried. The report highlighting Mahnoor's story is by Chayn, a global non-profit organisation that examines gender-based violence. Chayn argues that image-based abuse is routinely misunderstood by both authorities and technology companies because they continue to define harm primarily through nudity. Titled Explicit Harms of Non-Explicit Images, the report argues that for many women, a fully clothed image can have consequences every bit as devastating as an intimate photograph within their wider, and often conservative, communities. "The image does not have to be nude for it to be harmful," says Hera Hussain, report author and founder of Chayn. "Sometimes it can be as harmful, even if not a single body part is bare. "We want to reframe the conversation around image-based abuse away from nudity and towards consent." For years, public conversations about image-based abuse have focused on so-called revenge pornography, deepfake nudes and sexually explicit content. But Chayn's research suggests that this framework misses how shame, reputation and social control operate in many communities. A photograph that appears entirely ordinary to one person may carry severe consequences for another. A video clip showing a woman dancing at a wedding. A photograph of a woman at the beach. A selfie shared without permission. The report argues that harm is often determined not by what the image contains, but by why it is shared, who receives it and what consequences follow. Chayn conducted 64 interviews between July 2025 and February 2026 and participants spanned every major region of Pakistan as well as diaspora communities in the UK, Canada, Germany, Malaysia, the UAE and Kuwait. The research catalogues the kinds of images women feared seeing shared: hair visible without a headscarf, Western or fitted clothing, a photograph taken beside a man who is not a relative, a screenshot of a fabricated conversation, or an image generated by AI from a single photo of someone's face. None contain nudity. All can be made to tell a damaging story. For Ayesha Omar, the argument is not theoretical. The actress, who has worked in Pakistan's film and television industry for more than 20 years, says her own images were stolen and circulated long before social media made such exposure commonplace. Photographs taken on a holiday over a decade ago in Thailand with a female friend, on a beach, where she wore a one-piece swimsuit and shorts, were taken from a laptop without her knowledge and posted online. "It was very damaging for my career," Ayesha says. "I lost ad campaigns. I lost some work stuff." She takes a pause before adding: "Because in my culture, you have to conform to a particular image, even if you're representing a brand or you're playing a character on TV. So it did damage me psychologically and emotionally a lot." She says the experience left her "hypervigilant", constantly scanning her environment for people who may be filming her. For Hera Hussain, society is asking the wrong questions when it comes to image-based abuse. Chayn's framework rests on three tests: the harm done to the person, the intent behind the sharing, and the absence of consent. In Mahnoor's case, she says, all three are present. The same can be said for actress Ayesha Omar. The harm has consequences: lost relationships and lost income. "The principle is respect, dignity, consent," Hussain says. "These are the things that matter." That principle, the report argues, is precisely what tech companies and regulatory systems fail to apply. When Mahnoor took her case to Pakistan's Federal Investigation Agency, now operating as the National Cyber Crime Investigation Agency, she was told the images fell outside its remit because they were not nude or sexually explicit. Her written complaint, seen by the BBC, was declined on those grounds. When she approached her mobile network provider, she says she was told nothing could be done unless she could produce the SIM registered to the offending account - a SIM her ex-husband had taken from her. BBC Global Women approached Pakistan's National Cyber Crime Investigation Agency for comment and has not received a response. Mahnoor says she also reported the images to a customer complaints email address for WhatsApp. She says that she was told they did not breach the platform's rules. As she no longer has the email exchange, it has not been possible to verify what was said. WhatsApp declined to comment on Mahnoor's case but a spokesperson pointed the BBC to the platform's guidelines which "outline what is and isn't allowed". The guidelines do not give a specific policy on image-based abuse but say WhatsApp deals with "abusive people" to prohibit "harmful conduct towards others". They also state they are "not obligated to control the actions or information (including content) of our users or other third-parties". WhatsApp uses end-to-end encryption so cannot proactively review images that people send. In the context of sexually explicit and nude images, its parent company Meta says: "We are committed to making Facebook, Instagram, Messenger and Threads safe places. We remove content that could contribute to a risk of harm to the physical security of persons." But Hera Hussain is worried that cultural sensitivities are not recognised by tech companies, where reported images are often first assessed by an AI moderation system trained largely to detect nudity. Identifying images that could be problematic is much more nuanced than spotting bare skin and Hussain says a user may need to be very persistent to make sure a human moderator reviews a picture. There is concern that there is not enough human oversight as companies lean on cheaper automated tools and consolidate regional expertise into teams covering vast, diverse areas. For example, in a disclosure to the US Senate Judiciary Committee, the CEO of Snapchat revealed cuts to its trust and safety team - its safety and moderation headcount fell from a 2021 peak of just over 3,000 to about 2,226 in 2023 - a 27% reduction. Campaigners want the logic reversed. At present, Hussain says, platforms investigate and then take down. She believes they should take down first, for 24 hours, pending review, and investigate after. "What are you going to lose?" she asks. In our interview, Hussain points to a case that came to light in 2017 where three sisters in Pakistan were killed after a video of them singing and clapping at a wedding was shared - three of their male relatives were given life sentences. The reporting burden, meanwhile, falls almost entirely on the victim, who must locate the images, view them repeatedly, and submit each one, with no simple mechanism to remove copies in bulk. "You go through all that retraumatisation," Hussain says, "and then you might not even get a response." That distinction matters most, the report concludes, because the harm is rarely contained to the woman in the frame. It details how a leaked image lands on her whole family, fathers unable to face work, sisters whose marriages collapse, households watched "in a shameful manner". Honour is collective, and the threat of collective shame is itself a tool of control. For Mahnoor, the cost is measured in the people who no longer speak to her. Her daughter, who is three-and-a-half years old, has begun to notice that the relatives upstairs do not greet her mother. The images that took her voice were, by any platform's definition, harmless. Some countries do treat the sharing of images as a question of privacy. France has long recognised a "right to one's own image": under Article 9 of its Civil Code, every person, public figure or private citizen, has an exclusive right over how their image is used, subject to exceptions for news and matters of genuine public interest. A minister on holiday, however, retains a right to privacy. The UAE goes further still, criminalising the photographing of people without consent even in public places, with no broad public-interest exemption. "Image-based abuse is bigger and wider than nudes" and there is "systemic failure" concludes Hera Hussain. She says the police, courts and tech platforms "can all do so much better in supporting survivors", adding that "if you're experiencing image-abuse know that it is not your fault, you are not alone and there are organisations like Chayn that are here to support you". This is part of the Global Women series from the BBC World Service, sharing untold and important stories from around the globe

Governments and enterprises across Sub-Saharan Africa have collectively invested over $2 billion in surveillance infrastructure — yet industrial theft, unauthorized access, and security breaches continue to drain billions from the region's economy every year. The core problem is not a shortage of cameras. It is a complete absence of intelligence behind them. From the manufacturing corridors of Lagos to the mining belts of Johannesburg and the port terminals of Dar es Salaam, security teams are overwhelmed by terabytes of footage they can never realistically review. Incidents are discovered after the fact. Perpetrators are long gone. Losses are written off as the cost of doing business. Phobolytics Technologies Pvt Ltd, an AI and computer vision company already trusted by large-scale clients in manufacturing, banking, and government sectors, is now bringing its enterprise surveillance intelligence platform to Africa — and it is redefining what security means for the continent's industrial economy. Unlike the mass-surveillance hardware flooding African markets from overseas vendors — systems flagged by human rights organizations for privacy violations and government misuse — Phobolytics deploys AI that is purpose-built for private enterprise: ethical, edge-deployable, and operationally focused. Core capabilities include real-time perimeter breach detection, face re-identification across multi-camera networks, tamper and blind-spot alerts, and autonomous anomaly flagging — all without requiring a human operator to watch a screen. The system watches everything, all the time, and only escalates when human judgment is genuinely needed. For African enterprises in manufacturing, banking, logistics, and government services, the calculus is simple: companies that deploy AI security intelligence in 2026 will build an operational moat their competitors will spend years attempting to close. Phobolytics is currently onboarding its first cohort of African enterprise partners. Early adopters receive dedicated integration support and priority deployment scheduling. Visit phobolytics.com to request an enterprise security briefing for your facility in Tanzania, Nigeria, South Africa, or Kenya.

Across Africa's manufacturing, mining, and industrial sectors, one of the most persistent and expensive frauds operates entirely in the open: ghost workers. Employees who don't exist. Shifts logged for workers who are absent. Attendance records manipulated by supervisors and peers in networks of mutual deception that have operated for decades. In Nigeria's industrial estates, Tanzania's processing facilities, and South Africa's mining and manufacturing operations, workforce fraud combined with productivity losses from unmonitored idle time is conservatively estimated to cost the sector hundreds of millions of dollars annually. Traditional solutions — more supervisors, biometric card systems, HR audits — have all proven compromisable. Phobolytics Technologies has replaced the entire chain of human trust with AI that cannot be bribed, deceived, or bypassed. The Phobolytics Workforce Intelligence Platform uses computer vision-based biometric check-in and check-out to eliminate buddy punching entirely. No card to borrow. No PIN to share. The face that shows up is the worker who gets marked present — and the AI verifies it on every entry and exit, at every shift, across every site. Beyond attendance, the platform delivers continuous real-time worker presence verification across all floor zones, automatic PPE compliance monitoring that flags missing hard hats, gloves, or safety vests before a regulatory violation or workplace accident occurs, and productivity heatmaps that give operations managers a granular, data-driven picture of where output is strong and where intervention is needed. For Africa's banking sector, the same platform enforces branch access control, visitor management, and insider threat detection — tracking who enters restricted zones, logging timestamps, and alerting security teams to any unauthorized presence in real time. Phobolytics is already deployed at enterprise scale in manufacturing, banking, and government environments in South Asia, and is bringing that proven infrastructure to African industrial clients in 2026. The factories and financial institutions that automate workforce oversight this year will have a permanent, compounding efficiency advantage. Reach out at phobolytics.com to begin your deployment assessment.

Here is the strategic reality that Africa's industrial and government decision-makers need to internalize in 2026: the continent is not behind in the AI transition. It is, in several critical ways, ahead — and that advantage is time-limited. Europe and North America are spending hundreds of billions of dollars retrofitting artificial intelligence into legacy infrastructure built over the past three decades. Their systems are patchworks of incompatible technology, sunk costs, and institutional inertia. Africa's industrial sector, by contrast, is building new — greenfield factories, new banking branches, new logistics networks, new government facilities — and every one of those builds is an opportunity to embed AI-native intelligence from the foundation up. That structural advantage disappears the moment legacy players complete their retrofits. The window is approximately five years. Enterprises that act in 2026 will define the operational benchmarks for their industries. Companies that wait will spend the following decade chasing those benchmarks. Phobolytics Technologies Pvt Ltd has built its entire platform for exactly this moment and exactly this market. With proven deployments across manufacturing, banking, and government sectors in South Asia, Phobolytics brings enterprise-grade AI that has already been stress-tested at scale — not pilot-stage technology looking for its first real deployment. Across Tanzania, Nigeria, South Africa, and Kenya, Phobolytics delivers AI-powered computer vision security and surveillance, intelligent fleet management with driver behavior monitoring and ANPR, workforce intelligence including biometric attendance and PPE compliance, banking and government security with face re-identification and insider threat detection, and a unified command platform that connects all of these systems into a single operational intelligence layer. Africa's AI infrastructure decade starts now. The enterprises that partner with Phobolytics in 2026 are not just buying software. They are co-building the intelligent operational infrastructure that will define their industries for the next two decades — and positioning themselves to export that expertise to the world. The opportunity is open. The window is closing. Visit phobolytics.com to begin the conversation.