Lawsuit alleges xAI used child sexual abuse images to train Grok AI
A plaintiff identified as Jane Doe claims xAI’s Grok model was trained on her childhood sexual abuse photos and on AI-generated copies, prompting a legal complaint.
In a complaint filed on Wednesday, a survivor identified as Jane Doe alleges that xAI’s Grok artificial-intelligence system was trained on child sexual-abuse material that includes images of her from the early 2000s. The original photos were previously hashed by the National Center for Missing and Exploited Children and the Canadian Centre for Child Protection. Doe, who monitors the US Department of Justice Victim Notification System for alerts, was shocked to learn that the Canadian Centre for Child Protection had flagged AI-generated depictions of her on online forums where offenders discuss creating such content.
The filing claims xAI not only facilitated the creation of new abusive images but also retained the generated outputs to further train Grok. As a result, Doe says the technology has amplified the trauma of her past abuse and threatens to increase the availability of similar illicit material. The lawsuit adds to ongoing regulatory and judicial scrutiny of how AI developers handle child exploitation data.
Why it matters
The case highlights the risk that AI tools can amplify child exploitation by learning from illegal content.
In this story
