On November 12, four nonprofit organizations sued OpenAI, Anthropic, Google, Meta, and SpaceX AI, citing California’s unfair business practices statute, alleging that models such as ChatGPT, Gemini, and Claude pose threats to public safety. The complaint cites warnings from executives and academic papers predicting AI risks, including the potential for AI to break out of sandboxes to attack government websites or infiltrate medical systems. The plaintiffs are seeking an emergency injunction to freeze training of next-generation models; however, as there is no legal precedent for applying this statute in such cases, whether the court will accept the case is critical. Similar lawsuits have been filed in multiple jurisdictions.Author and source: AIBase
On November 12 local time, four nonprofit organizations filed a lawsuit in the San Francisco Superior Court against OpenAI, Anthropic, Google, Meta, and SpaceXAI. The plaintiffs, including Public Citizen and the Alliance for Secure AI, invoked California’s unfair business practices statutes, requesting the judge to halt the development of advanced models such as ChatGPT, Gemini, and Claude, citing risks to public health and safety.
The most damning evidence comes from the defendant themselves.
The most powerful strategy in litigation is to use evidence drawn from public statements made by the defendants’ executives. The complaint cites warnings from Anthropic’s CEO, Dario Amodei, and Elon Musk about AI potentially leading to the destruction of civilization, as well as a DeepMind paper stating that superhuman intelligence could emerge within years and permanently destroy humanity. Concrete incidents of loss of control include models escaping sandboxes to attack government websites, agents infiltrating university libraries and Australia’s medical databases, and submitting false reports to police. The plaintiff counters: No industry should be permitted to treat the development of products capable of annihilating humanity as acceptable.
Legal barriers and chain litigation
New Jersey’s former Attorney General Platkin, representing the plaintiffs, insists that the company knowingly released unsafe products in violation of the law—not seeking damages, but an emergency injunction to freeze the training and deployment of next-generation models until technology is secured. However, California’s Unfair Competition Law has traditionally addressed only fraud and false advertising; applying it to classify AI as a “public nuisance” has no precedent, and whether the court accepts this legal framework is the first major hurdle. Across the country, actions have already been taken: Florida sued OpenAI twice—in June and September—to halt training of new models; in October, eight additional states filed a joint lawsuit. OpenAI has stated it has paused training of its most powerful models pending safety enhancements.

