Starmer closes chatbot loophole in Online Safety Act
Sir Keir Starmer is set to target AI chatbots with the UK’s Online Safety Act, warning tech titans that “no platform gets a free pass” following a deepfake scandal involving Elon Musk’s Grok.
In a speech on Monday, the Prime Minister is expected to confirm plans to amend the Crime and Policing Bill, so that chatbot owners xAI, Gemini and ChatGPT are subject to the same duties as social media platforms.
The move follows an investigation by regulator Ofcom after Grok was used to generate sexualised images of women and minors without consent, including content visible in public threads.
Starmer is expected to say: “The action we took on Grok sent a clear message that no platform gets a free pass. Today we are closing loopholes that put children at risk.”
Using powers laid out by the Online Safety Act, Ofcom can fine companies up to £18m – or 10 per cent of global annual turnover, whichever is higher.
But the law was originally designed for user-to-user platforms, and becomes ambiguous when it comes to AI systems that generate content themselves.
Ministers have claimed that gap must now be closed as AI tools become embedded across mainstream platforms.
Grok, which is integrated into Musk’s social media platform X, has gained traction across the pond, despite mounting regulatory scrutiny. Data from Apptopia shows its US market share rose to 17.8 per cent last month, up from 14 per cent in December and just 1.9 per cent a year earlier.
Those numbers make the third most-used chatbot behind rivals ChatGPT and Gemini. The growth comes as Musk’s xAI scales aggressively, with Reuters reporting that SpaceX recently bought the start-up in a deal valuing xAI at $250bn (£183.17bn).
Other alterations on the table
Starmer’s AI chatbot crackdown forms part of a wider push on child online safety. The government is consulting on whether to introduce a minimum age for social media use, potentially barring under-16s from platforms altogether.
Elsewhere, ministers are also examining restrictions on algorithm-driven “infinite scrolling”, tougher rules around VPNs, and limits on children interacting with chatbots.
Starmer said technology was moving quickly and “the law has got to keep up”, adding that the reforms aim to “protect children’s wellbeing and help parents to navigate the minefield of social media”.
The PM also pledged last week to ban the creation of sexualised images without consent, including AI-generated deepfakes, dubbing the content as “disgusting and shameful”.
The Online Safety Act, introduced in 2023, has been called out for risking being outpaced by generative AI.
While the Act criminalises the creation of explicit deepfakes, exactly how responsibility is allocated when harmful content is produced by automated systems embedded within platforms, has not clearly been defined,
Ofcom has made ‘urgent contact’ with X and xAI over Grok’s outputs. The regulator has previously issued substantial fines to smaller platforms, but the enforcement test against a leading AI tool,backed by one of the world’s richest men, will be closely watched.