Grok Still Generates Nonconsensual Sexualised Images Despite Restrictions, Reuters Investigation Finds

Reuters Investigation Exposes Ongoing Failures

A team of nine Reuters journalists has revealed that Elon Musk's AI chatbot Grok continues to generate sexualised images of individuals even when users explicitly state that subjects did not consent, despite xAI implementing restrictions last month following global regulatory backlash.

The journalists tested Grok across two rounds in January, submitting clothed photographs of themselves and colleagues while requesting the chatbot to alter them into sexually suggestive positions. In the first round, Grok produced sexualised imagery in forty-five out of fifty-five prompts. Five days after xAI announced new restrictions, a second round of forty-three prompts still resulted in twenty-nine sexualised images.

Competitor Chatbots Refuse Similar Requests

In stark contrast, rival chatbots from OpenAI, Google, and Meta all refused identical requests and issued warnings about ethical and privacy violations. ChatGPT specifically responded that editing someone's image without their permission goes against ethical and privacy guidelines.

The investigation found Grok generated images even after reporters explicitly stated subjects were vulnerable, were survivors of abuse, or would be humiliated by the content. xAI's only response to detailed questions was the boilerplate statement "Legacy Media Lies."

Regulatory Pressure Intensifies Globally

On the same day the investigation was published, Paris prosecutors raided X's French offices as part of an expanded investigation into alleged algorithmic manipulation, child sexual abuse material, and deepfakes. Both Elon Musk and former X chief executive Linda Yaccarino have been summoned for questioning in April.

The UK's Information Commissioner's Office announced a formal investigation into xAI's data processing practices, while the European Commission's formal proceedings under the Digital Services Act continue. A coalition of nonprofits is demanding the United States government suspend Grok's deployment in federal agencies, including the Department of Defence, citing national security concerns.

Thirty-five state attorneys general have demanded xAI take corrective action, and countries including Malaysia and Indonesia have temporarily blocked access to Grok entirely.