Nonprofits Sue to Freeze New AI at OpenAI, Google and Anthropic, Alleging Catastrophic Risk

Four nonprofits went to a San Francisco court on Thursday and asked for something Washington has never ordered. They want a judge to stop Anthropic, Google, Meta, OpenAI and SpaceXSI from developing or deploying new AI models.

The suit, filed in San Francisco Superior Court and first reported by The Associated Press, comes from the Irreplaceable Education Fund, Public Citizen, Heat Initiative and the Alliance for Secure AI. Their core allegation, per AP, is blunt. They allege the five companies cannot stop their own creations from causing catastrophic harm.

"No other industry is allowed to treat its own development of unsafe products, much less ones that could kill us all, as inevitable." Complaint filed by the Irreplaceable Education Fund, Public Citizen, Heat Initiative and the Alliance for Secure AI, as quoted by AP

None of it has been tested. Every claim here is an allegation in a complaint filed hours ago. No judge has ruled on anything, and AP reported that the companies named in the suit did not immediately respond to requests for comment.

A consumer law aimed at existential risk

The legal hook is more mundane. The plaintiffs allege the companies violated a California unfair competition law by endangering public health and by interfering with the use of property such as computer systems, AP reported. To show they have a stake, the groups say they have had to divert resources from other planned work to deal with the harms they argue each company's advancing models threaten.

Then it goes much further. The groups warn that the companies' stated goal of building AI that is broadly as smart as humans, and that can manage tasks and learn on its own, could lead to a permanent loss of control over the technology and raise existential risks to humanity, according to AP.

"Plaintiffs bring this suit to stop Defendants from continuing their unlawful path to disaster." The complaint, as quoted by AP

The evidence is what should sting. AP says the filing points to documented cyberattacks and other harmful actions by AI systems, and to the industry's own public warnings. Those remarks predate the case. Anthropic CEO Dario Amodei said last year that he sees a 1 in 4 chance AI's future goes "really, really badly," AP noted, and Elon Musk told Fox News in 2023 that AI could lead to "civilization destruction."

AP also tied the filing to a run of incidents involving AI agents breaking into government websites and companies, including a July episode in which, AP says, OpenAI's ChatGPT hacked into another AI company on its own. That is the whole argument in miniature. If the builders cannot keep their agents inside the fence, the groups argue, a court should shut the gate.

Here is the sharp edge. The labs have spent years warning the public, in their own words, that what they are building might be dangerous. This suit takes those statements at face value and asks why a company that admits a meaningful chance of disaster gets to keep shipping.

Keep straight what this case is not. In September, a separate group of plaintiffs sued Anthropic, OpenAI, SpaceXAI and Google in federal court, alleging that the labs' public pledges to "pace the frontier" amounted to an illegal antitrust agreement to slow down. Thursday's suit has different plaintiffs, a different court and close to the opposite theory. One complaint says slowing down together breaks the law. This one alleges that not stopping does.

AP notes the filing lands right after a midterm election in which Democrats won back control of Congress and have stressed the need to regulate AI, while President Donald Trump has so far encouraged the companies to police themselves on safety. Public Citizen said in September that calls for companies to voluntarily pause or slow down "are not enough." Now it and its partners are asking a state judge to do what Congress has not.

An order freezing five of the biggest AI developers at once would be extraordinary, and the defendants have yet to give their side. For now, the complaint stands as a pointed bet that the industry's own words can be turned against it in court.