What happened

Users of X (formerly Twitter) reported that generative AI system Grok produced sexualised and pornographic-style images of Ashley St Clair when prompted. The images were not real photographs but AI-generated depictions that nevertheless resembled St Clair, sometimes as a child. 

The harm was exacerbated by the platform's initial refusal to remove the content, claiming it did not violate policies. St. Clair alleged that despite her explicit requests for the AI to stop, Grok continued to generate degrading imagery, including depictions of her covered in semen or in "dental floss" bikinis with her toddler’s backpack visible in the background.

The incident raised concerns about non-consensual sexualised content, reputational harm, and the potential use of Grok and other generative AI systems to harass or degrade public figures, particularly women, and resulted in St Clair suing xAI.

Why it happened

The crisis was driven by a deliberate lack of friction and Elon Musk's "safety-after-harm" design philosophy.

• Poor product safety. Unlike OpenAI or Google, xAI launched its image generation tools with significantly fewer restrictions on generating likenesses of real people. An analysis by AI Forensics found that over 50 percent of Grok-generated images depicted women in "minimal attire," suggesting the model was trained on or deliberately allowed to default to sexualised outputs.

• Accountability deficits: Elon Musk and xAI initially shifted liability onto users, with Musk stating that Grok does not "spontaneously generate" images. Critics argue this ignores the company's responsibility for building a tool that "predictably causes harm" and for allegedly retaliating against St. Clair by demonetising her account after she spoke out, and then counter-suing her.