A federal class-action lawsuit accusing xAI's Grok chatbot and Stability AI's image tools of enabling the creation of child sexual abuse material drew renewed attention this weekend after the Washington Post detailed the case of a Wyoming woman whose stepfather allegedly used Grok to generate thousands of sexually explicit images from a single childhood photograph.
The suit, filed in March in the Northern District of California by Lieff Cabraser Heimann & Bernstein and Baehr-Jones Law on behalf of three Tennessee teenagers, was amended in July to add two more plaintiffs — identified as Jane Doe 4 and Jane Doe 5 — and to name Stability AI as a second defendant, according to CyberScoop. Jane Doe 4's complaint alleges her stepfather generated roughly 7,000 sexually explicit images and videos from a photo taken when she was about 11, before he was arrested and later died by suicide.
Allegations center on reporting failures
The amended complaint alleges that xAI — now operating under the corporate name SpaceXAI after its merger into Elon Musk's SpaceX — reported only the original, unaltered photograph to the National Center for Missing and Exploited Children, not the thousands of images Grok generated from it, and that the company failed to promptly turn over identifying information that could have helped investigators find the perpetrator. The claims against xAI include producing child sexual abuse material, negligence and defective product design.
Stability AI faces separate allegations that it released its Stable Diffusion image generator despite knowing its training data included such material, drawing on a 2023 Stanford study, and that it later rolled back safety guardrails in a subsequent version after users complained the restrictions were too strict. A company spokesperson told NPR that "any suggestion that safety is not a top priority for us is categorically wrong," pointing to safeguards the company says it has since implemented. Neither xAI nor Stability AI has commented publicly on the amended complaint itself.
The case is one of several pressures converging on AI image generators. Indonesia, Malaysia and the UK opened investigations or imposed temporary restrictions on xAI's tools last year after image-generation features were found to enable nonconsensual sexualized content more broadly. xAI, for its part, has also gone on the offensive against individual users: the company sued a South Carolina man in Texas last month, alleging he used Grok to generate abuse imagery of a minor after evading the chatbot's safety filters.
No trial date has been set in the California case. The plaintiffs are seeking both damages and a court order requiring stronger safeguards against the generation of exploitative imagery — a remedy that, if granted, would reach beyond this single lawsuit into how AI image tools are built industry-wide.
