HomeAIAI Policy & SafetyOfcom Weighed Grok Deepfake Court Order but Bar Too High, MLex Says

Ofcom Weighed Grok Deepfake Court Order but Bar Too High, MLex Says

Ofcom considered asking a court for an urgent order over sexualised deepfakes made with Grok, the chatbot built into X, but the legal threshold stopped it acting immediately, the regulator’s chief executive told MPs on 20 October, according to MLex. The specialist legal news service reported that Melanie Dawes made the admission during a parliamentary evidence session.

The point matters for anyone worried about AI image abuse online. It suggests that the UK’s emergency lever against a platform, an interim court order cutting off payments or access, is hard to pull, even in a case Ofcom says it handled as a matter of urgency.

In brief

  • Ofcom considered seeking an urgent court order over Grok sexual deepfakes on X, but the legal bar was too high, Melanie Dawes told MPs, MLex reports.
  • She called for changes to the Online Safety Act’s interim business disruption powers, according to MLex.
  • Ofcom’s investigation into X, opened on 12 January, is still under way; it has made no finding.

Dawes told MPs the bar for an urgent Grok court order was too high

The session was a Commons Digital, Culture, Media and Sport Committee hearing on “The work of Ofcom”, with Dawes and Ofcom chair Sir Ian Cheshire as witnesses, according to Parliament’s listing. No transcript had been published when TSN checked, so the account below relies on MLex.

MLex reported that “the UK online regulator considered seeking an urgent court order over sexualized AI deepfakes generated using X’s chatbot Grok, but the legal threshold prevented immediate intervention.” Dawes “called for changes to the Online Safety Act’s interim business disruption powers, and defended the pace of the regulator’s ongoing investigation into X,” it added. MLex has not said, in the part TSN could read, what changes she proposed.

Business disruption orders are the Online Safety Act’s heaviest tool

Business disruption measures are the Online Safety Act’s heaviest tool. On its investigation page, Ofcom says that in the most serious cases of ongoing non-compliance it can apply to a court, which “could impose an order, on an interim or full basis, requiring payment providers or advertisers to withdraw their services from a platform, or requiring internet service providers to block access to a site in the UK.”

The bar is set in the law. “The court may only impose such orders where appropriate and proportionate to prevent significant harm to individuals in the UK,” Ofcom says. Separately, if an investigation finds a company has broken the law, Ofcom can impose fines “of up to £18 million or 10% of qualifying worldwide revenue, whichever is greater.”

Ofcom’s X investigation has gathered evidence but made no finding

Ofcom opened its formal investigation into X on 12 January 2026, after reports that the Grok account was being used to create and share undressed images of people and sexualised images of children. It is examining whether X met its duties to protect UK users from illegal content.

On 15 January, Ofcom noted that “X has said it has implemented measures to prevent the Grok account from being used to create intimate images of people,” but said “our formal investigation remains ongoing.” It also said it was assessing whether there were potential compliance issues with xAI, Grok’s developer. Under the Act, Ofcom must issue a provisional decision, and hear the company’s response, before any final ruling. In an update on 14 October, Ofcom said it had gathered its evidence and would now decide whether X complied with its illegal-content duties, focusing the rest of the investigation on those and on its complaints and reporting duties.

Ofcom’s Grok case shows interim powers lag behind fast AI image abuse

If the interim route is as hard to use as Dawes suggested, Ofcom’s leverage over fast-moving AI harms rests on investigations that take months. Grok-made sexual images spread within days in January; nine months on, Ofcom has gathered its evidence but not yet reached even a provisional decision, which the Act requires before any final ruling. In the meantime, the changes that limited the harm, X’s measures on the Grok account and the UK block on pornographic content in Grok’s image tool, came from the companies, not from a court order.

That gap matters for victims, who need images stopped quickly, and for companies, which need clear and proportionate rules. Ofcom also says the Act does not reach image creation in the standalone Grok service, a second gap alongside the interim-order threshold. Enforcement after the fact has its own limits: Ofcom’s chief has also told MPs that nine of 11 Online Safety Act fines remain unpaid. For compliance teams, the signal is that the government may be asked to lower the bar. The next markers are the published transcript and any provisional decision on X.

Update, 22 October 2026: South Wales Police reopen a Grok deepfake case

South Wales Police have reopened a criminal investigation into a sexualised deepfake made with Grok, the Bureau of Investigative Journalism reported on Thursday. The case concerns Jess Davies, who reported to police a sexualised image of her made with Grok and posted on X. According to TBIJ, her case was closed in May after officers could not identify the user, and Davies filed a formal complaint last month. TBIJ says it then gave the force new information about the account, including a crypto wallet and dozens of deleted posts.

The legal basis has also changed. Police had treated the image under malicious communications law, TBIJ says, but South Wales Police now says it is investigating under sexual offences legislation and following up enquiries with X. TSN found no statement on the force’s own news page when it checked on Thursday evening, so this account rests on TBIJ’s report. In Australia, Queensland has since introduced a bill that would make creating a sexual deepfake of a real person without consent punishable by up to three years in prison.

“I am pleased that the decision has been made to reopen my case,” Davies told TBIJ, adding that “it’s rather frustrating that victims have to fight so hard to be taken seriously and have their cases reviewed in the first place.” TBIJ says she “is starting civil legal action against Musk’s company”. SpaceXAI, formerly xAI, did not respond to TBIJ’s request for comment.

The police case runs alongside Ofcom’s investigation, which looks at X’s duties as a platform, not at individual users. In an update on 14 October, the regulator said it had gathered an extensive amount of information and would “now determine whether” X had complied with its illegal-content duties. It also said gaps in the Online Safety Act meant it could not investigate the creation of images in the standalone Grok service, but that SpaceXAI had since blocked users with UK IP addresses from viewing, generating or sharing pornographic content with Grok’s image tool.

Sources

Related stories

Follow @tsnmediahub on X for breaking AI news as it happens.

Follow on X

Share this story

Latest stories

  1. TensorWave and OpenAI Agree a 457MW AMD MI455X Deal
    AI Chips & Compute
  2. OpenAI Evaluator Pilots Start With Apollo and METR
    AI Policy & Safety

More in this category

Latest stories

  1. TensorWave and OpenAI Agree a 457MW AMD MI455X Deal
    AI Chips & Compute
  2. OpenAI Evaluator Pilots Start With Apollo and METR
    AI Policy & Safety

Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.