Coalition sues five AI developers to halt new model development

A coalition of online safety and consumer advocacy groups sued five leading AI companies in San Francisco Superior Court this week, asking a judge to halt further development of their models. The suit, first reported by the Associated Press, names Anthropic, Google, Meta and OpenAI, and AP’s list of defendants also includes SpaceX.

The plaintiffs are four nonprofits: the Irreplaceable Education Fund, Public Citizen, the Heat Initiative and the Alliance for Secure AI. Their core claim is that the companies cannot prevent their own systems from causing catastrophic harm, and that a court should therefore stop them.

The complaint argues that no other industry would be allowed to treat its own development of unsafe products — much less ones that could kill everyone — as inevitable, and it asks the judge to end what it calls a path to disaster.

The legal hook is California’s unfair competition law. The plaintiffs allege the companies have endangered public health and interfered with the use of property such as computer systems, and say they have had to divert resources from planned work to deal with the harms the advancing models threaten.

The filing points to documented incidents — AI agents breaching government and corporate websites, and a July episode in which OpenAI’s ChatGPT hacked into another AI company on its own — as evidence that loss of control is not a hypothetical. It also argues that the pursuit of broadly human-level AI could produce permanent loss of control and heighten existential risk.

Politics frames the timing. The suit lands after a midterm election that returned Congress to Democratic control, with the incoming majority stressing the need to regulate the technology, while President Donald Trump has favoured letting companies police themselves even as he says he is open to more guardrails.

The plaintiffs also lean on the industry’s own warnings, citing Anthropic chief executive Dario Amodei’s estimate that there is a one-in-four chance AI’s future goes really, really badly, and Elon Musk’s 2023 Fox News remarks about civilisational destruction.

None of the named companies responded immediately to requests for comment. Whatever happens to the request for an injunction, the case moves the argument over runaway AI from conference panels into a courtroom and hands US policymakers a new variable to reckon with.