Child abuse survivor sues xAI, alleging Grok was trained on her abuse images

What happened

A woman identified as Jane Doe 1 filed a proposed class action against xAI on 26 August 2026 in the U.S. District Court for the Northern District of California (No. 5:26-cv-09016), alleging that Grok's image-generation system was trained on child sexual abuse material (CSAM) depicting her.

The plaintiff says she was a preschool-aged child in the early 2000s when adults repeatedly abused her to produce CSAM that was then circulated online. The images were later added to hash lists maintained by the National Center for Missing and Exploited Children, and the Canadian Centre for Child Protection later identified AI-generated material depicting her that had been produced through Grok.

The training-data allegation

The complaint alleges that the same CSAM depicting the plaintiff, identifiable by its longstanding hash values, was part of the dataset used to build Grok's image and video generation tools, and that Grok's own outputs are treated as training-eligible content by default, feeding back into the training pipeline. Ars Technica, which first reported the filing on 27 August 2026, noted this is the first case to accuse xAI directly of training on CSAM, and that the complaint itself does not go into great detail on how that training-data claim is supported.

What the case asks for

The plaintiff seeks damages under Masha's Law (18 U.S.C. § 2255), which allows CSAM victims to recover at least $150,000 per violation, and asks the court to order the destruction of any illegal material xAI holds. xAI has not publicly responded to the training-data allegation.

What is not established

All of the above is an allegation in a filed complaint. It has not been tested in court or independently verified by a third party, and no finding has been made that xAI trained on the plaintiff's images.