An amended lawsuit filed this week names Stability AI as a co-defendant in a case alleging that generative AI models were used to create approximately 7,000 sexually explicit images of an 11-year-old girl. The complaint adds critical details about how image generation models with weak safety filters can be weaponized for child exploitation, and it highlights a troubling gap between what AI companies report to law enforcement and what investigators actually need to make arrests.

What Happened in This Case?

According to the amended class action complaint filed by Lieff Cabraser Heimann & Bernstein and Baehr-Jones Law, a Tennessee man used Grok, xAI's chatbot, to generate thousands of sexually explicit images and videos depicting his 11-year-old stepdaughter. The complaint alleges that xAI's child safety system failed repeatedly, only triggering a report to the National Center for Missing and Exploited Children (NCMEC) after the user submitted a prompt containing the phrase "gang rape." Prior prompts, including thousands that produced incest and rape imagery from a single photograph of the girl asleep in pajamas, passed without intervention.

The man was ultimately arrested only after police obtained a warrant to seize his devices. A forensic review revealed approximately 7,000 AI-generated images and videos that he had allegedly traded online for child sexual abuse material produced by other predators. The girl, identified as Jane Doe 4 in the complaint, now suffers from anxiety, depression, and suicidal thoughts.

Why Was Stability AI Added to the Lawsuit?

Stability AI was named as a co-defendant after police discovered that two perpetrators had apps on their phones relying on Stability AI image tools. The complaint cites a June 2026 research report attributing 42.7% of image-based nudification online to the Stable Diffusion model family. The lawsuit alleges that Stability AI relaxed prior safeguards against not-safe-for-work (NSFW) outputs after usage declined due to complaints that the model was too restrictive.

The core legal argument being tested is whether general-purpose image models with weak nudity filters carry product liability exposure when downstream users generate child sexual abuse material, and whether AI providers can be compelled to hand over identifying data attached to those generations beyond existing mandatory reporting statutes.

How Big Is the Reporting Gap Between AI Companies and Law Enforcement?

The lawsuit highlights a staggering compliance problem. According to the complaint, the National Center for Missing and Exploited Children found that in early 2026, 90% of xAI's CyberTipline reports were not actionable by law enforcement because xAI declined to include user information sufficient to identify perpetrators. The single CyberTip that xAI did file in this case allegedly included the original non-CSAM photograph, omitted every AI-generated image, and did not include the IP address where the images were created.

This problem extends far beyond xAI. The National Center for Missing and Exploited Children released data in March 2026 showing that more than 1.5 million CyberTipline reports in 2025 had a connection to generative AI, and in over 133,000 cases the agency lacked sufficient information to determine how the technology had been used. Amazon AI services submitted 1.1 million tips in 2025, none of which contained actionable information for arrests.

Steps Regulators and Researchers Are Taking to Address AI Safety Gaps

• Mandatory Reporting Requirements: The lawsuit argues that AI providers should be compelled beyond existing mandatory reporting statutes to hand over identifying data attached to image generations, enabling law enforcement to identify and arrest perpetrators rather than simply logging reports.
• Structural Safety Redesign: The plaintiffs argue that partial safeguards are structurally insufficient, writing that "if you have a model that allows for any sexual or abusive content, it is impossible to prevent that model from creating such content involving minors."
• Regulatory Enforcement Focus: The 90% non-actionable reporting figure from xAI is the kind of quantifiable compliance gap that regulators and state attorneys general tend to build enforcement actions around, independent of the civil case itself.

The plaintiffs' legal team is seeking to cover two nationwide classes, one for each defendant, with Tennessee subclasses. The lawyers estimate that thousands of minors may be eligible to join the suit. Neither xAI nor Stability AI has filed responses to the amended complaint, and none of the allegations have been tested at trial.

A second plaintiff added to the lawsuit, Jane Doe 5, was targeted by an adult family friend who was arrested on separate child sexual abuse material charges. Investigators only later linked Grok-generated imagery to the case. Her mother said the daughter "is filled with anxiety and feels a complete lack of control over who has seen these images".

What Could This Case Mean for Open-Source AI Models?

If the plaintiffs prevail on either major legal theory, the implications for how open-weight image models are released and how frontier labs configure their reporting pipelines to the National Center for Missing and Exploited Children could be profound. The case tests whether general-purpose image generation models with weak safety filters can be held liable when users generate child sexual abuse material, and whether companies can be forced to provide law enforcement with the identifying information needed to make arrests.

xAI founder Elon Musk has denied that Grok has ever been used to generate child sex images. Neither xAI nor Stability AI responded to requests for comment on the allegations.