“No other industry”: Safety groups sue OpenAI and other AI giants to stop new models
Four advocacy groups have sued five leading AI companies in California, asking a court to stop the development and deployment of new AI models over alleged safety risks. OpenAI, Google, Meta and Anthropic are among the companies named, according to the Associated Press, which first reported the filing.
The plaintiffs argue the companies are advancing systems without being able to prevent potentially severe harm. Those are allegations in a new lawsuit, not findings by a court. The AP report does not describe an order requiring the companies to stop building or releasing models.
“No other industry is allowed to treat its own development of unsafe products” as inevitable, the complaint says, according to the AP. The news agency said the companies did not immediately respond to its requests for comment.
Anthropic is among the AI companies named in the lawsuit, the Associated Press reported. Image: Anthropic
What the groups are asking for
The case was filed in San Francisco Superior Court by the Irreplaceable Education Fund, Public Citizen, Heat Initiative and Alliance for Secure AI, the AP reports. It says the nonprofits are seeking an order that would prevent the defendants from developing or deploying new models.
The groups allege violations of a California unfair competition law. They argue that risks to public health and computer systems have forced them to divert resources from other work, according to the report.
The requested remedy is broader than a dispute over one chatbot feature. But a request is not the same as a ruling: the companies can challenge the allegations, and the report does not establish that any existing service is being shut down.
OpenAI has previously published a framework for assessing risks from advanced AI. Image: OpenAI
AI companies have published safety plans
The case arrives as developers are publishing their own approaches to advanced-model risk. In May, OpenAI described a frontier governance framework covering areas including cyber offence, harmful manipulation and loss of control. The company said its existing preparedness framework remained the foundation for managing serious risks.
Separately, Anthropic’s published roadmap describes planned policy proposals on independent evaluators and the pace of AI development; these proposals are not binding requirements. These documents were published before the lawsuit and are not responses to the complaint.
The central question for the case is whether the plaintiffs can persuade the court to grant the restrictions they seek. The AP did not report an immediate response from the companies, and their positions in the litigation remain to be seen.
