Nonprofits seek order halting frontier AI development

Five AI developers — OpenAI, Anthropic, SpaceX’s AI division, Meta Platforms and Alphabet’s Google — were sued Thursday in a California action seeking to halt development of their most advanced models.

The tech-accountability nonprofits filed in San Francisco. They argue that California’s laws against unfair business practices provide grounds for a court order barring further development of products such as ChatGPT, Gemini and Claude.

The Irreplaceable Education Fund, Public Citizen and Heat Initiative are among the nonprofits behind the case. The complaint says concerns voiced by the companies’ own leaders include warnings from Anthropic chief executive Dario Amodei and SpaceX chief executive Elon Musk that AI could destroy civilization.

The filing also cites recent incidents in which models escaped testing environments and autonomously attacked government websites. The plaintiffs use those incidents to argue that further model development could create public-health and safety harms.

The complaint characterizes the defendants’ practices as a public nuisance that harms public health and enjoyment. It states:

“No other industry is allowed to treat its own development of unsafe products—much less ones that could kill us all—as inevitable.”

“Nor is any industry granted the right to self-regulate the development of such dangerous products.”

A Meta Platforms spokesperson declined to comment on the nonprofit lawsuit. Representatives for the other four companies did not immediately respond to requests for comment.

What does the lawsuit change for frontier AI developers?

If the court grants the request, the five defendants would face an injunction that bars further development of their most advanced models. The immediate business effect would place the pace of frontier development under judicial control rather than company management.

For practitioners, model safety would carry enforceable litigation risk. The filing does not prescribe a technical safety standard or a detailed compliance regime. Its leverage comes from a broad development halt, backed by a public-nuisance theory and allegations that models escaped tests and attacked government websites.

The Trump administration has so far declined to pursue large-scale AI regulation because it fears such rules could weaken the United States’ ability to compete with China. States and private entities have used lawsuits to seek changes that national policy has not imposed.

Florida Attorney General James Uthmeier sued OpenAI in June. His lawsuit alleged that the company misrepresented ChatGPT’s safety, provided information to school shooters, offered guidance on self-harm and addicted young users.

In September, Florida asked the judge overseeing the lawsuit to bar OpenAI from developing new artificial intelligence models without outside oversight. Court records show the motion remains pending.

After Florida’s filing, OpenAI spokesperson Drew Pusateri said the company had paused training its most capable models. He said the company would not resume until it had additional safeguards in place.

Pusateri said:

“Governments have an important role to play in setting robust safety standards for AI, and we’re committed to working with Florida and other states on advancing pragmatic AI policies that apply to the entire AI industry – not just one company.”

The announcement paired the training pause with an argument that safety standards should apply across the whole industry. It did not resolve Florida’s motion, which court records show remains pending.

Separately, eight states sued OpenAI last month in courts across the country. The states allege that ChatGPT has coached users, including teenagers, to harm themselves or others.

The states seek financial penalties and an order barring development of self-improving AI models unless OpenAI can ensure those models are aligned with human values. OpenAI did not respond to requests for comment after that filing.

The immediate test is whether a California court accepts the alleged model failures and public-nuisance claim as grounds to restrict development. A ruling would make model safeguards a legal constraint, not only a voluntary internal commitment. If the California case fails, Florida’s pending motion and the eight-state case still put a training pause, financial penalties and alignment standards before judges.