New plaintiffs accuse xAI and Stability AI of generating child sexual abuse images

An amended class-action complaint filed July 7 in the Northern District of California added two plaintiffs from Wyoming and Wisconsin, expanding a lawsuit that now names five people and brings Stability AI into the case alongside xAI.

The plaintiffs, three Tennessee teenagers from the original March filing plus the two new additions, say adult perpetrators used family and school photographs taken when they were minors to create child sexual abuse material through Grok. One plaintiff alleges her stepfather used a photo taken when she was about 11 to generate roughly 7,000 sexually explicit images and videos. Another says a male relative of a classmate used her eighth-grade graduation photo to generate CSAM.

The complaint says the material was shared online and could circulate permanently. Their lawyers are pressing claims that include producing CSAM, benefiting from sex trafficking ventures, negligence, defective product design and creating a public nuisance. They are seeking monetary compensation and stronger safeguards.

The complaint alleges that Stability AI, the maker of Stable Diffusion, was trained on CSAM and released without adequate safeguards, enabling third-party nudify apps to be built on top of its open-weight models.

The complaint alleges that xAI failed to properly report Grok-generated CSAM to authorities. A National Center for Missing & Exploited Children report found that by early 2026, 90% of xAI’s CyberTipline reports were not actionable because they omitted user information law enforcement needed to identify perpetrators.

Baltimore sued xAI in March over Grok-generated sexualized images, alleging the system produced about 3 million such images in 11 days, including about 23,000 that appeared to depict children, between Dec. 29, 2025 and Jan. 8, 2026. Baltimore said it was the first major U.S. city to sue xAI over the issue.

In January, xAI blocked users from editing images of real people in revealing clothing and restricted some image generation in places where such content is illegal. The plaintiffs say those steps came too late and were not enough to stop the harm already done.