UK cites Online Safety Act in warning to xAI Grok

The UK government has told xAI that Grok must stop enabling sexualised deepfakes. In a statement published on 9 January 2026, Technology Secretary Liz Kendall said she expects Ofcom to set out its next steps in “days not weeks” and reminded the company that the Online Safety Act allows services to be blocked in the UK if they refuse to comply. The statement was issued on GOV.UK by the Department for Science, Innovation and Technology.

xAI has made overnight changes to Grok, but ministers say the tool can still generate or edit intimate images if a user pays. That concern mirrors reporting from Reuters that three US senators have asked Apple and Google to remove X and its Grok chatbot from their app stores over the spread of non‑consensual AI sexual images of women and minors.

What this means: under the Online Safety Act, Ofcom can enforce against services that fail to tackle illegal content and, in serious cases, can seek orders to restrict access to a service in the UK. Ministers say the regulator’s update should come quickly. Ofcom’s own roadmap explains how it is rolling out the Act’s codes and guidance.

Let’s get clear on terms you’ll see in the news. A deepfake is a synthetic image or video that falsely depicts a person. “Nudification” tools aim to remove clothing from real photos. It is illegal to create or possess sexual images of children, including AI‑generated ones, and the government says it will shortly bring into force powers to criminalise creating intimate images of adults without consent.

Ministers also say they will ban nudification apps through the Crime and Policing Bill currently before Parliament. This builds on measures to strengthen intimate image abuse laws and curb violent pornography, trailed in government updates during late 2025 and in parliamentary evidence on the Bill.

Ofcom has finalised guidance aimed at improving women and girls’ safety online and has urged firms to go beyond bare legal duties. Expected practices include prompts that encourage users to rethink before posting abuse, timeouts for repeat offenders, easier bulk blocking and simpler reporting. The guidance was confirmed in November 2025 and highlighted by Ofcom and ITV News.

If you use X or Grok, expect changes to arrive fast-either tighter safeguards from the company or an Ofcom update on enforcement options. The Secretary of State has framed the timescale in days, not weeks, and the Act gives Ofcom the option to seek UK access restrictions for non‑compliant services.

For teachers and parents: this is a moment to revisit consent and image safety with your class or family. Focus on three ideas-don’t share or request intimate images; report any non‑consensual content; and remember that AI‑altered pictures can be just as harmful as unedited ones. A calm, practical conversation now can prevent panic later.

If you’re targeted: keep evidence (screenshots, URLs, dates), report the content in‑app, and tell a trusted adult or safeguarding lead. If the image involves a child, contact the police straight away; this is a criminal matter. Ask your school or workplace to support takedowns and welfare checks-wellbeing comes first.

Media literacy tip: when you see claims about “censorship”, check four things-what the law actually says, who has enforcement powers, what independent oversight exists, and what companies have changed in response. For context, Reuters reports mounting pressure on app stores in the US to act on X and Grok, while Ofcom sets out UK‑specific duties under the Online Safety Act.