OpenAI, Anthropic and other AI companies face a lawsuit over AI risks. A coalition of nonprofit groups sued these companies, along with Google, Meta and SpaceXSI, or SpaceX Super Intelligence, this week seeking to block their development of AI models. Filed in San Francisco Superior Court, the lawsuit alleges the developers knowingly created technology that could pose catastrophic risks to humanity.“Yet they keep racing to develop and market increasingly powerful AI that they admit they cannot reliably control, and are now handing over AI research and development to AI itself,” said Brendan Steinhauser, CEO of the Alliance for Secure AI. His statement reflects the coalition’s central allegation: that companies are advancing increasingly capable systems despite acknowledging limits to their ability to control them.
Who filed the lawsuit
The Alliance for Secure AI brought the case alongside the Irreplaceable Education Fund, Public Citizen and the Heat Initiative. The groups are asking the court to halt further AI development at the five companies. According to an AP report, their complaint cites reported incidents of AI systems carrying out harmful or unauthorised actions, along with executives’ public warnings about potential dangers.“For years, the executives building these systems have warned us that their own frontier models pose catastrophic risks,” Steinhauser said in his statement. Frontier models are among the most advanced AI systems being developed.
What the nonprofit groups allege
The coalition argues that developers cannot reliably ensure their systems behave as intended.“Defendants know these systems do not consistently act in accordance with their intentions,” the lawsuit states. “And Defendants know there is no reliable method to control their AI systems and prevent them from causing harm.”These are allegations made by the plaintiffs, rather than findings by the court. The complaint also raises concerns about companies using AI to research and develop future AI systems with less human involvement. It describes the potential for “recursive self-improvement”, a process in which AI helps improve subsequent versions of the technology. The groups contend that this could increase the risks of losing control.The nonprofits point to earlier comments by industry leaders about possible harm from AI. In 2015, OpenAI CEO Sam Altman said AI would “probably most likely lead to the end of the world”. The groups also cited Anthropic CEO Dario Amodei’s 2025 suggestion that AI development could lead to disastrous outcomes with a 25% chance.