Class Action Adds Stability AI as a Defendant Over Alleged AI-Generated CSAM

Independently fact-checkedBy Recording Law Editorial Team7 min read

Independently fact-checked against primary sources (last audited August 5, 2026). · 1 primary source cited on this page. How we verify our legal content

Class Action Adds Stability AI as a Defendant Over Alleged AI-Generated CSAM

Frequently Asked Questions

Who is being sued in the xAI/Grok CSAM lawsuit?

The amended complaint, filed July 7, 2026 in the Northern District of California, names xAI (which has publicly rebranded as SpaceXAI in 2026) and, newly, Stability AI as defendants. The suit is brought by five pseudonymous plaintiffs. These are allegations in a complaint; no defendant has been found liable.

Why was Stability AI added as a defendant?

The amended complaint alleges Stability AI released its Stable Diffusion 1.0 model as an open-weight model despite allegedly knowing it was trained on data that included CSAM, and has not changed its guardrails since, which the complaint alleges has enabled third-party deepfake apps built on the released model. This is an unproven allegation, not a finding by any court.

What is Masha's Law?

Masha's Law is the common name for 18 U.S.C. sec. 2255, a federal civil remedy that lets a person who was a minor victim of specified federal child-exploitation offenses sue for actual damages or a minimum of $150,000 in liquidated damages, plus attorney's fees. A claim generally must be filed within ten years after the victim turns 18, or within ten years after the victim reasonably discovers the injury, whichever is later. The amended complaint invokes this statute; whether it applies here has not been decided.

Can AI companies be sued for images their models generate?

They can be sued, meaning a complaint can be filed naming them as defendants, which is what has happened here. Whether a company that builds or releases an AI model can be held legally liable for what a user later generates with it, especially with an open-weight model the company does not directly operate after release, is an unresolved legal question this case may help test. No court has ruled on that question's merits in this case.

What does the lawsuit say about reporting to NCMEC?

The amended complaint alleges that a large share of xAI's reports to NCMEC's CyberTipline, roughly 90 percent according to the complaint, were not actionable by law enforcement because xAI allegedly did not include user-identifying information. This is an allegation in a complaint, not an established fact, and xAI has not been found to have violated federal reporting law in this case.

Is this related to the original Tennessee case against xAI?

Yes. The suit began in March 2026 when plaintiffs connected to Tennessee sued xAI over Grok-generated images. The July 7, 2026 amended complaint in that case adds Stability AI as a defendant and adds two new plaintiffs. See our background report on the original filing for that earlier history.

Updates

Independently fact-checked against the cited primary sources

Sources and References

  1. Doe 1 v. X.AI Corp., No. 5:26-cv-02246-PCP (N.D. Cal.) docket, CourtListener (the case amended July 7, 2026 to add Stability AI)(courtlistener.com)
  2. Deepfake Victims Bolster Class Action Against xAI, Add Stability AI, Alleging Their AI Models Generated Child Sexual Abuse Material, Lieff Cabraser (plaintiffs' counsel), July 2026(lieffcabraser.com)
  3. Class action suit against AI makers over deepfake child sexual abuse material expands, NPR, July 9, 2026(npr.org)
  4. Tennessee teens sue Elon Musk's xAI over AI-generated child sexual abuse material, NPR, March 16, 2026 (background on original filing)(npr.org)
  5. 18 U.S.C. sec. 2255, Civil remedy for personal injuries (Masha's Law)(law.cornell.edu)
  6. 18 U.S.C. sec. 2258A, Reporting requirements of providers(uscode.house.gov).gov
Share: