X Moves to Curb Grok’s AI ‘Undressing’ of Real People After Backlash

Elon Musk’s social media platform X has announced new restrictions on its AI tool Grok, saying it will no longer allow users to edit images of real people to depict them in revealing clothing in countries where such content is illegal. The move follows mounting outrage over the creation and spread of sexualised AI deepfakes.

In a statement posted on X, the company said it had introduced “technological measures” to prevent Grok from being used to alter images of real individuals in sexualised ways. The change comes after weeks of criticism from campaigners, victims, and regulators who warned that the tool was enabling serious abuse.

The UK government welcomed the announcement. A spokesperson described it as “vindication” of calls for X to rein in Grok, while media regulator Ofcom called it a “welcome development”. However, Ofcom stressed that its investigation into whether X breached UK law is still ongoing.

Campaigners and victims say the decision comes too late to undo the harm already caused.

“We are working round the clock to understand what went wrong and what’s being done to fix it,” Ofcom said, adding that enforcement action remains on the table if X is found to have failed in its legal duties.

Technology Secretary Liz Kendall also welcomed the move but said she expects Ofcom to “fully and robustly establish the facts” through its investigation.

Journalist and campaigner Jess Davies, whose image was among those altered using Grok, said the changes were a “positive step” but criticised the platform for allowing the abuse to happen at all.

“It’s sobering to think how many women, including myself, have been targeted, and how many more victims of AI abuse are being created,” she told the BBC, describing X’s overall response as “really pathetic”.

“They’re just trying to do as little as possible within very loose legal guidelines,” she added.

Dr Daisy Dixon, a philosophy lecturer at Cardiff University who was also targeted, said the images left her feeling “shocked”, “humiliated” and fearful for her safety. While she called X’s reversal a “battle win” for campaigners, she stressed that the damage had already been done.

“The abuse should never have happened,” she said. “Many women are now left dealing with lasting harm. Our relationship with our own bodies has been hijacked and distorted against our will.”

Andrea Simon, director of the End Violence Against Women Coalition, said the decision showed the power of pressure from victims, campaigners and governments. But she warned that tech companies must go further.

“Given the evolving nature of AI-generated harm, platforms must be required to take proactive, preventative action,” she said.

Political Pressure Mounts

X’s announcement came just hours after California’s attorney general confirmed the state was investigating the spread of sexualised AI deepfakes, including images of children generated by AI tools.

The company said it would now geoblock users in certain countries from generating images of real people in bikinis, underwear or similar attire through Grok. It also reiterated that only paying subscribers can access image-editing features, arguing this would make abusers easier to identify and hold accountable.

Musk has said Grok’s NSFW settings are meant to allow only limited nudity involving fictional adults, in line with US film ratings, with rules varying by country. Earlier this week, however, he dismissed critics as enemies of free speech and reposted AI-generated images of UK Prime Minister Sir Keir Starmer in a bikini—actions that drew further criticism.

Experts remain sceptical about how effective X’s new safeguards will be. Questions remain over how Grok will determine whether an image shows a real person and whether users will be able to bypass regional blocks using tools like VPNs.

Ofcom has warned that if X is found to be in breach of UK law and fails to comply, it could seek a court order to block the platform in the UK altogether.

Policy researcher Riana Pfefferkorn said the safeguards should have been implemented as soon as the abuse emerged and criticised Musk’s conduct for undermining the seriousness of the issue.

“It would help if he stopped doing things like reposting AI images of political leaders in bikinis,” she said.