Elon Musk's social media platform X has announced restrictions preventing its Grok AI chatbot from editing images of real people into revealing clothing, following mounting pressure from regulators and governments over the proliferation of non-consensual sexual deepfakes.

The company said it had "implemented technological measures to prevent the Grok account from allowing the editing of images of real people in revealing clothing such as bikinis", with the restriction applying to all users including paid subscribers. X added it would geoblock the ability to generate such images in jurisdictions where it is illegal.

The move came hours after California's attorney general Rob Bonta launched an investigation into the spread of sexualised AI deepfakes generated by Grok, including images of children. "The avalanche of reports detailing the non-consensual, sexually explicit material that xAI has produced and posted online in recent weeks is shocking," Bonta said in a statement. California governor Gavin Newsom called xAI's actions "vile", writing on X that the decision to "create and host a breeding ground for predators" was unacceptable.

British Prime Minister Sir Keir Starmer had earlier described the situation as "disgusting and shameful", telling MPs at Prime Minister's Questions that X was "acting to ensure full compliance with UK law". The government has pledged to introduce legislation this week making it a criminal offence to create non-consensual intimate images. "If so, that is welcome, but we're not going to back down, and they must act," Starmer said.

UK media regulator Ofcom launched a formal investigation into X on Monday over concerns the platform may have breached the Online Safety Act. An Ofcom spokesperson called the latest restrictions "a welcome development" but said its investigation "remains ongoing". If X is found non-compliant, Ofcom could impose fines of up to 10 per cent of worldwide revenue or £18 million, or seek court orders to block the site in Britain entirely.

Musk defended the AI tool on Wednesday, posting that he was "not aware of any naked underage images generated by Grok. Literally zero." He added that Grok "does not spontaneously generate images" but responds only to user requests, and is programmed to "refuse to produce anything illegal" according to local laws.

The controversy prompted calls from three Democratic US senators for Apple and Google to remove X and Grok from their app stores. Last week, X had restricted image generation to paid subscribers only, though technology secretary Liz Kendall dismissed this as "monetising abuse". Several countries including Malaysia and Indonesia have implemented bans or legal action against the platform.

Jonathan Lewis, UK managing director of X, said restricting features to paid subscribers would add "an extra layer of protection by linking a feature to identifiable paid subscribers who also can be held accountable". However, questions remain about enforcement mechanisms and whether the restrictions can be circumvented using tools such as virtual private networks.

---