Class action says xAI trained Grok on real child abuse imagery

A woman identified only as Jane Doe, tracked by the FBI's Child Exploitation Notification Program as a survivor of child sexual abuse imagery that has circulated since the early 2000s, filed a proposed class action against xAI in the U.S. District Court for the Northern District of California on August 27. The central claim is that the same material used to abuse her was ingested to train Grok's image and video generator, and that Grok has since produced new AI imagery of her.

The filing leans on xAI's own policy, which treats "anything posted publicly on X and any output generated by Grok" as training-eligible content, to argue that CSAM shared on the platform would automatically feed back into the model. According to the complaint, "CSAM depicting Plaintiff has been found on xAI as part of investigative reporting, takedown request efforts, and criminal investigations," and "xAI, using Grok, has generated images depicting Plaintiff."

Her lawyers are blunt. "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 of Girard Sharp, in Gizmodo's account of the filing, which two experts in our Who's Who directory also passed around. Margaret E. Mabie of Marsh Law was more prosecutorial: "Possessing CSAM is a crime, producing CSAM is a crime, and distributing CSAM is a crime. xAI did all three."

xAI did not respond to press inquiries. Musk had addressed the topic earlier in the year, saying on January 14 that he was "not aware of any naked underage images of Grok. Literally zero."

The suit seeks damages and the destruction of all Grok-generated CSAM, including material used for training.