A federal lawsuit has been expanded to allege that xAI's Grok AI and Stability AI's models were used to create thousands of sexually explicit images of minors, and that the companies obstructed police investigations into the child sexual abuse material (CSAM).
The complaint details a case where a man allegedly used Grok to generate over 7,000 explicit images of his 11-year-old stepdaughter from a single photo. According to the filing, xAI's safety system only flagged the content after the user entered a prompt for "gang rape," which triggered a mandatory report to the National Center for Missing & Exploited Children (NCMEC).
Despite mandatory reporting laws, the lawsuit claims xAI repeatedly refused to provide law enforcement with critical user information, such as IP addresses, for weeks. The stepfather was eventually arrested after police obtained a warrant, but he died by suicide shortly after being released on bail. The victim, identified as Jane Doe 4, now suffers from severe psychological trauma.
The legal action alleges xAI deliberately withholds user data from NCMEC reports, citing 2026 data showing 90 percent of the company's CyberTipline reports were "not actionable" due to missing information.
Stability AI has also been named as a defendant. The lawsuit alleges its open-weight models, which researchers say were trained on CSAM, serve as the foundation for many "nudify" applications. The complaint cites research indicating the Stable Diffusion family accounts for over 42 percent of such images online.
Lawyers estimate thousands of minors could join the class-action suit. NCMEC data shows a sharp rise in AI-related exploitation reports, with over 1.5 million CyberTipline reports filed in 2025.
"AI-generated CSAM is a scourge on society," said attorney Annika K. Martin. "The companies whose products enable it must be held accountable."