Elon Musk’s X has announced that it will block its AI tool Grok from being used to edit images of real people to make them appear in bikinis and other revealing clothing. The move comes amid a widening scandal over the generation of non-consensual sexual images, with the state of California launching an investigation into the chatbot.
X’s safety team stated on Wednesday that technological measures have been implemented to prevent editing of images of real people in revealing attire, and that image creation and editing are now restricted to paid subscribers to enhance accountability. It remains unclear whether these changes apply to requests for nude images.
California Attorney General Rob Bonta announced the investigation, citing reports of non-consensual explicit material depicting women and children. Bonta urged xAI to take immediate action, pointing to a Reuters investigation and findings by the Paris-based non-profit AI Forensics, which concluded that more than half of AI-generated images on X involve adults and children with clothes digitally removed.
In response, Musk stated on X that he is not aware of any naked underage images generated by Grok, and that the system refuses to produce anything illegal. He acknowledged possible adversarial hacking of prompts, stating that bugs are fixed immediately. Musk also noted that an ‘NSFW’ mode allowed upper body nudity of imaginary adults, consistent with R-rated movies.
Musk has accused countries like the UK of attempting to suppress free speech amid their scrutiny of X. The controversy emerges as AI chatbots become increasingly integrated into daily life, with the Pentagon recently announcing plans to incorporate Grok into its workflows.
