X restricts Grok AI from altering images of real people after backlash

Elon Musk's artificial intelligence model Grok has introduced technical safeguards to prevent users from editing images of real people to make them appear in revealing clothing, such as bikinis or underwear, following international backlash and growing legal scrutiny over AI-generated sexualised content.

The new restrictions apply to all Grok users, including paid subscribers. While the system continues to allow upper body nudity of imaginary adult humans under its NSFW settings, it now explicitly bans similar depictions involving real individuals, says the BBC.

The changes come amid mounting regulatory pressure. California's Attorney General recently opened an investigation into the spread of deepfake images, including those involving children. In the United Kingdom, media regulator Ofcom is examining whether X, the social media platform owned by Musk, is complying with relevant laws. Authorities in Malaysia and Indonesia have already moved to ban the Grok tool.

The issue has also taken on political significance in Britain. Prime Minister Sir Keir Starmer, who was previously targeted by AI-generated images shared by Musk, warned that X could lose its ability to self-regulate if it failed to address the problem. Although Starmer welcomed the new safeguards, critics and policy researchers said the platform acted too slowly and argued that Musk's own reposting of AI-generated images of politicians weakened its credibility.

Experts have raised concerns about enforcement, questioning how the platform will reliably distinguish between real and imaginary people in images. The controversy has prompted some UK schools and local councils to consider withdrawing from X, citing safety and content moderation concerns.

X has said that restricting advanced image editing tools to paid users provides an additional layer of protection, as it allows the company to identify and hold accountable those who violate laws or platform rules.