US lawsuit accuses Anthropic, OpenAI, Google of colluding to slow AI

A civil lawsuit filed in a United States federal court on Friday accuses four leading artificial intelligence companies of colluding to slow the pace of AI development, alleging that their public statements amounted to an illegal agreement between competitors.

The complaint, lodged in the Northern District Court of California, names Anthropic, OpenAI, SpaceXAI and Google. It contends that remarks made earlier this month by their senior figures constituted an unlawful business arrangement under US antitrust law, and specifically violates the Sherman Act, which bars rivals from joining forces to restrict competition.

The dispute traces back to Anthropic Chief Executive Officer Dario Amodei, who publicly urged "industry-wide coordination" to "pace the frontier" of artificial intelligence advancement. SpaceXAI head Elon Musk, OpenAI CEO Sam Altman and Google DeepMind co-founder Demis Hassabis swiftly voiced agreement with that call, according to the filing.

The suit has been brought by four plaintiffs: lawyer Cheyenne Hunt, Florida attorneys Charles Buist and Nick Spetsas, and California resident Christine Bullock. They have signalled their intention to seek class-action status on behalf of others affected by any corporate decisions to slow development.

Nick Rowley, one of the lawyers representing the plaintiffs, said the action is aimed at preventing private arrangements from sidelining regulatory oversight. He argued that the case was filed so that "private self-serving agreements between the world's most powerful 'for profit' technology companies" do not allow AI to "quickly spin out of human control."

Rowley added that humanity deserves firm safeguards against extinction-level threats, and that the rule of law should be set transparently and lawfully by government, with accountability to the public.

Plaintiff Cheyenne Hunt announced her involvement on social media platform X, saying she was taking action against Anthropic, OpenAI, X and Google because the public deserves genuine AI safety standards rather than arrangements made by an unaccountable group of billionaires. She alleged that the companies had been meeting privately since July to plan a response to mounting public pressure for AI regulation, and that Amodei's essay calling for a slowdown was followed within hours by public agreement from Musk and Altman.

Hunt disputed the framing of the move as a safety breakthrough, characterising it instead as four corporations agreeing not to compete on the very safety standards regulators were preparing to require. She said the timing coincided with lawmakers gaining momentum on binding AI rules, and criticised the arrangement as one written, enforced and revisable by the companies themselves.

She further claimed the public pledge gave House Speaker Mike Johnson cover to adjourn the House of Representatives shortly before binding legislation on the subject could be passed. Hunt argued that decisions about technologies posing potential extinction risks should not rest with four chief executives weighing profit margins, but with government, under real transparency.

The plaintiffs are seeking to halt what they describe as private deals that bypass legislative scrutiny.