xAI sued for allegedly training Grok on child sex abuse material

A lawsuit filed Wednesday accuses Elon Musk's xAI of training its Grok models on real and AI-generated child sex abuse material (CSAM), marking the first case to make that specific claim against the company. The plaintiff, identified as Jane Doe, was preschool-age in the early 2000s when she was raped to produce CSAM that was sold online. Her images have since been hashed by the National Center for Missing and Exploited Children and the Canadian Centre for Child Protection. The CCCP notified Doe it had identified AI-generated CSAM on xAI depicting her, re-traumatizing the plaintiff. Her lawyers allege that Doe's images — which carry well-known hash values — "was part of the dataset xAI used to build Grok's image and video generating capabilities." The complaint also alleges offenders discussed creating AI-generated CSAM of Doe on online forums, and that xAI stores Grok's outputs and uses them for further training. The complaint provides limited detail on the training claim, and Ars notes there's no indication xAI used a previously scrubbed dataset that contained CSAM. Some Grok users have been arrested as regulators probe the issue.