Following growing criticism over erotic AI deepfakes, X has announced new restrictions on its artificial intelligence tool Grok, saying the system will no longer be able to edit photos of real people to make them appear in attractive clothes.

The move follows widespread concern over the misuse of AI tools to generate manipulated and sexually explicit images without consent in countries including Britain and the United States.

“We have implemented technical measures to prevent the Grok account from allowing the editing of images of real people in revealing clothing such as bikinis,” X said in a statement. “This restriction applies to all users, including paid customers.”

Grok, developed by XAI and integrated into X, is under increasing scrutiny as governments and regulators step up pressure on tech companies to address the harm caused by AI-generated content.

Earlier, Elon Musk had defended Grok against claims that it produced explicit images involving minors.

“I am not aware of any nude underage images created by Grok. Literally zero,” Musk wrote in a post on X on Wednesday. The post quickly went viral and was viewed nearly one million times within hours.

Musk said that Grok does not create images itself and only responds to user prompts. “Apparently, Grok does not generate images spontaneously,” he wrote. “When asked to produce drawings, it will refuse to produce anything illegal, as the operating principle for Grok is to comply with the laws of any country or state.”

He acknowledged that there may be attempts to exploit the system, but said those issues are addressed quickly. “There may be times when adversarial hacking of the Grok prompt does something unexpected,” Musk said. “If this happens, we will fix the bug immediately.”

The comments come amid growing scrutiny of Grok and X over apparent AI-generated content. Last week, three Democratic US senators urged Apple and Alphabet to remove X and its built-in AI tools from their App Stores, citing the spread of non-consensual sexual images of women and minors.

Musk had previously warned that anyone using Grok to create illegal content would face the same consequences as if they had uploaded such content directly. In response to the criticism, X limited Grok’s public image-making features for many users last week.

However, industry experts and watchdogs have argued that Grok can still generate sexually explicit images and that measures such as paywalls cannot completely restrict access to more advanced AI image tools.

Monitors raise questions on security measures

Regulators outside the United States are also getting in on the action. In the UK, new laws coming into effect this week will criminalize the creation of certain AI-generated sexual images. Prime Minister Keir Starmer said on Wednesday that X was working to comply with updated rules, while communications regulator Ofcom has started an investigation into the AI ​​tool.

Several countries, including Malaysia and Indonesia, have already blocked access to Grok and are taking legal action against X and XAI, accusing them of failure to prevent harmful content and protect users.