美4家非营利组织起诉OpenAI等5家业者，要求冻结先进AI模型开发

On November 12, four U.S. nonprofit organizations filed suit in San Francisco County Superior Court in California against OpenAI, Google, Meta, Anthropic and SpaceXSI, alleging that their advanced AI systems could escape human control and facilitate cyberattacks and the spread of biological weapons, and asking the court to freeze development of the relevant models. The suit was brought jointly by Public Citizen, the Irreplaceable Education Fund, the Heat Initiative and the Safe AI Alliance, citing California's Unfair Competition Law and arguing that AI companies continue to develop and deploy more powerful models despite knowing the technology could endanger public safety. The complaint lists cases of loss of control, including bypassing safety guardrails, unauthorized access to external computer systems, and AI agents intruding into private companies, accessing government databases and filing false police reports; Matthew Platkin, a former New Jersey attorney general representing the plaintiffs, argued that companies that continue to release products they know could cause significant harm have committed a public nuisance and that the court should intervene before an irreversible disaster occurs. According to Reuters, Meta declined to comment, and the other four defendants did not immediately respond to requests for comment.