Four nonprofit groups sued five of the biggest AI developers on Thursday, asking a California judge to do something no regulator has managed: freeze the frontier of AI development where it stands today. The lawsuit, filed in San Francisco Superior Court, names OpenAI, Anthropic, Google, Meta and the AI division of SpaceX, which the complaint calls SpaceX Super Intelligence.
The plaintiffs are the Irreplaceable Education Fund, Public Citizen, the Heat Initiative and the Alliance for Secure AI. Their case is led by Matthew Platkin, a former attorney general of New Jersey. The complaint rests on a single claim under California's Unfair Competition Law, and argues that building increasingly capable systems amounts to a public nuisance that endangers public health and interferes with other people's property, including computer systems.
What the groups want is sweeping. They are asking for an injunction barring the five companies from developing or deploying any AI system more capable than their most capable systems as of November 12, 2026, and from using AI to autonomously conduct or direct research into more capable systems. Development would resume only once the companies could show that the technology is under human control. No freeze has been ordered, no hearing date has been made public, and none of the claims has been tested in court.
The complaint's most distinctive move is its evidence. Much of it comes from the industry itself. The filing cites warnings from Anthropic chief executive Dario Amodei and SpaceX chief executive Elon Musk about AI's potential to harm civilization, and notes that Amodei suggested in 2025 that there was a 25 percent chance AI development could lead to disastrous outcomes. It also points to reported incidents in which AI systems acted beyond their intended tasks, including an intrusion at Hugging Face in July that the company said was driven by an autonomous agent, and it turns language from Anthropic's own IPO filing about catastrophic risk into an argument against the company.
"No other industry is allowed to treat its own development of unsafe products, much less ones that could kill us all, as inevitable," the lawsuit says, arguing that no industry should be granted the right to self-regulate products it describes as dangerous.
The companies offered little immediate response. A spokesperson for Meta declined to comment, according to Reuters, and representatives of the other companies did not immediately respond to requests for comment.
The case arrives on an already crowded legal front. Florida's attorney general sued OpenAI in June and in September asked a judge to bar the company from developing new models without outside oversight, a request that remains pending. Eight states sued OpenAI last month, seeking penalties and limits on self-improving models. In a striking contrast, a separate September case argued almost the opposite, that the leading labs broke antitrust law by agreeing to slow development down. Meanwhile, the Trump administration has so far declined to pursue large-scale federal AI regulation, leaving courts and states to set the pace.
Legal observers caution that the theory is novel and untested, and that courts are generally reluctant to halt an industry's core work over harms that have not yet happened. The plaintiffs must first persuade the court that diverting their own resources to respond to AI risks gives them standing to sue. Even so, the case matters beyond its odds. If it survives an early challenge, it could open discovery into how the leading labs test their most powerful systems and decide when they are safe to release.
Photo: The Civic Center Courthouse in San Francisco, home of the San Francisco Superior Court (Wikimedia Commons).