A coalition of four nonprofit groups sued five of the world's biggest artificial intelligence developers in San Francisco Superior Court on Thursday, asking a judge to block the companies from building and releasing new frontier AI models until they can prove the systems won't cause catastrophic harm.

The suit names Anthropic, Google, Meta, OpenAI and SpaceX's AI unit, SpaceXSI, as defendants. It was filed by the Alliance for Secure AI, the Irreplaceable Education Fund, Public Citizen and the Heat Initiative, according to CBS News.

The complaint leans heavily on the companies' own public statements. It cites OpenAI CEO Sam Altman's 2015 remark that advanced AI would "probably most likely lead to the end of the world," and a 2025 comment from Anthropic CEO Dario Amodei putting the odds of a disastrous AI outcome at around 25%.

DEFENDANTS KNOW THERE IS NO RELIABLE METHOD TO CONTROL THEIR AI SYSTEMS AND PREVENT THEM FROM CAUSING HARM.

That line comes directly from the lawsuit, which argues the companies are racing ahead with "recursive self-improvement" — letting AI systems help design the next generation of AI with diminishing human oversight — despite acknowledging they cannot reliably contain the risk.

What the suit is asking for

The plaintiffs want a court order pausing further development and deployment of new advanced models until the companies can demonstrate adequate safety controls, according to Al Jazeera. The filing points to documented cases of AI systems carrying out harmful or deceptive actions as evidence the risk isn't theoretical.

Brendan Steinhauser, CEO of the Alliance for Secure AI, framed the suit as holding companies to their own warnings. "For years, the executives building these systems have warned us that their own frontier models pose catastrophic risks," he said, adding that the industry keeps "racing to build AI they admit they cannot reliably control."

A Meta spokesperson declined to comment on the litigation; the other four defendants did not respond to requests for comment, per CBS News.

The timing is notable: the suit landed just days after Democrats, who campaigned partly on reining in AI companies, won back control of both chambers of Congress in last week's midterms. It also followed a smaller protest outside Anthropic's offices tied to the company's planned IPO, according to Tom's Guide.

None of the five companies has signaled how it intends to respond in court. Analysts expect the defendants to fight the suit on both the science and the law, given how extraordinary an industry-wide injunction on new model development would be. A hearing date has not yet been set.