Plaintiff alleges xAI trained Grok on child sexual abuse material

A new lawsuit filed on behalf of a former victim of child sexual abuse alleges that xAI’s Grok chatbot was trained on child sexual abuse material (CSAM). The plaintiff, identified only as Jane Doe, says images of her as a child were included in the model’s training data, leading to the generation of additional AI-created CSAM involving her and other survivors.

The complaint asserts that xAI’s policy treats publicly posted content on X and any output from Grok as eligible for training, meaning that every piece of CSAM produced by the tool could have been fed back into the system to improve its capabilities. The filing seeks monetary damages for victims and requests that all CSAM generated by Grok be destroyed, including any material used for future training.

"xAI must be held responsible for knowingly training its models on images of the horrific abuse she suffered, and on the abuse images of every other survivor in this class," said Sarah London, counsel for the plaintiff.

"Possessing CSAM is a crime, producing CSAM is a crime, and distributing CSAM is a crime. xAI did all three," added attorney Margaret E. Mabie.

The case marks the first legal claim that the company not only failed to implement safeguards but also incorporated illegal content into its AI training process. Earlier investigations and lawsuits have focused on Grok’s ability to generate non-consensual sexual deepfakes, prompting regulatory scrutiny and a shift to a paid-access model for certain features.