Elon Musk’s AI model Grok will no longer allow users to edit images of real people into revealing clothing, following global concern over sexualised AI deepfakes. The restriction, which applies to all users, including paid subscribers, was announced on X (formerly Twitter) hours after California’s top prosecutor began probing AI-generated sexualised images, including of children.

X said users attempting to generate such images will be blocked according to local laws. Grok will still allow NSFW (not safe for work) edits of imaginary adults consistent with R-rated films, but not real people. The move follows bans in Malaysia and Indonesia and criticism from UK authorities, including Britain’s media regulator, Ofcom, and public figures like UK Prime Minister Sir Keir Starmer.

Experts still question how X will enforce the rules and verify whether images involve real people, while some say safeguards should have been implemented sooner. Musk’s prior posts, including AI-generated images of Starmer in a bikini, have drawn further criticism, highlighting ongoing concerns about AI abuse and accountability.