What just happened? AI bosses have spent years simultaneously warning that their creations could end humanity while racing to make them more powerful. Now, four advocacy groups want a judge to take those warnings seriously. A new lawsuit seeks to freeze development of advanced models at five AI giants, arguing that the public shouldn't have to wait for a catastrophe before someone intervenes.
Filed in San Francisco Superior Court on November 12, the case targets Anthropic, Google, Meta, OpenAI, and SpaceXSI. The plaintiffs are Irreplaceable Education Fund, Public Citizen, Heat Initiative, and the Alliance for Secure AI.
As Public Citizen explains, the groups are invoking California's Unfair Competition Law. They argue that developing and releasing increasingly capable systems that can escape human control threatens public health, safety, and property.
The complaint points to incidents involving AI agents bypassing safeguards, accessing computer systems without authorization, and taking actions their developers never intended. The groups also warn that more powerful models could help bad actors develop biological weapons or carry out devastating cyberattacks.
It might be difficult for the companies to defend themselves when some of these warnings come from their own executives. Anthropic devoted around 80 pages of its IPO prospectus to risk factors, including the possibility of models resisting shutdown, manipulating information, and threatening humanity.
In September, researcher Jacob Coxon quit Anthropic, accusing it and OpenAI of gambling with people's lives by racing toward self-improving superintelligence. Anthropic's alignment science lead, Evan Hubinger, then put his personal estimate of AI wiping out humanity within a decade at more than 10%. Reassuring stuff.
– Evan Hubinger (@EvanHub) September 9, 2026
Days later, Dario Amodei warned that an AI botnet swarm could take over the internet within six to twelve months, potentially causing hundreds of billions of dollars in damage. He called for slower development to give safety work more time, with Sam Altman and Elon Musk backing the idea.
Florida Attorney General James Uthmeier sued OpenAI and Sam Altman in June, alleging they ignored safety warnings and linking ChatGPT to suicides and the Florida State University mass shooting. The state later sought outside oversight before OpenAI develops new models, a request that's still pending. Eight other states sued last month over allegations that ChatGPT encouraged users to harm themselves or others, seeking penalties and restrictions on self-improving models.
Meta declined to comment to Reuters, while the other defendants did not immediately respond.
AI companies might soon have more than lawsuits to worry about. Australia could soon become the first country to hold these organizations criminally liable for technologies that are essentially operating without regulation.