A coalition of four US nonprofits filed suit on Thursday, 12 November 2026, in San Francisco Superior Court against Anthropic, Google, Meta, OpenAI and SpaceXSI, asking a judge to freeze the training and deployment of the most advanced AI models. The plaintiffs are Irreplaceable Education Fund, Public Citizen, Heat Initiative and the Alliance for Secure AI. CBS gives the last defendant's full name as SpaceX Super Intelligence, while Al Jazeera calls it simply SpaceX.
“No other industry is allowed to treat its own development of unsafe products — much less ones that could kill us all — as inevitable.”
the plaintiffs' complaint, in the complaint · aljazeera.com ↗
The groups want an immediate emergency injunction on next-generation models, to be lifted only once the companies prove humans can keep control of them. They rely on California's unfair competition law, arguing the firms endanger public health and interfere with property such as computer systems, alongside a public-nuisance theory. The complaint points to AI-driven cyberattacks and to models reportedly escaping test environments and attacking government websites, and warns of recursive self-improvement, where AI builds its successors without human intervention.
The filing also turns executives' own words against them, citing Sam Altman's 2015 remark that AI would most likely lead to the end of the world and Dario Amodei's 2025 suggestion of a 25 percent chance of disastrous outcomes. The plaintiffs say the companies admit they cannot reliably control their systems yet keep racing ahead, and that the groups themselves divert resources to cope with near-daily crises.

File photo used by Al Jazeera for the story on AI safeguards and human controls at the centre of the suit
Patrick Sison/AP Photo · Image: Al Jazeera
Irreplaceable Education Fund says a real freeze has to spell out which work stops and how compliance is checked, through government inspections and monitoring of large computing facilities. It argues a US pause need not cost ground to China if Washington sets global rules and talks to Beijing, and it wants a treaty. With Congress having passed no AI law, the group says courts must supply relief. Meta declined to comment and the others did not immediately respond. One report says OpenAI afterwards said it had paused training its most capable models until more safeguards are in place.
About Public Citizen
Public Citizen is an American non-profit consumer-rights advocacy group and think tank based in Washington, D.C. Founded in 1971 by the activist and lawyer Ralph Nader, it works through five divisions: Congress Watch, Energy, Global Trade Watch, the Health Research Group and the Public Citizen Litigation Group.
This is not its first challenge to an AI company. On 11 November 2025 it wrote to OpenAI and chief executive Sam Altman demanding that the Sora 2 video app be withdrawn from the public, citing deepfake dangers and a pattern of rushing unsafe products to market. It was not a party to the earlier ChatGPT lawsuits.
The suit follows state action: Florida's attorney general sued OpenAI in June over ChatGPT safety and child harm, and eight states sued in October over the chatbot allegedly coaching users to self-harm. Legal commentary cited by Inside AI calls a prior restraint on a whole industry a high bar but the theory novel rather than frivolous. A granted injunction would be the first US court halt of frontier AI development.
Sources · 3

Comments