Four AI giants face an antitrust lawsuit claiming they coordinated a development slowdown, putting private safety deals and consumer competition under federal court scrutiny.
Four paid subscribers filed the case against Anthropic, OpenAI, Google and SpaceXAI in the U.S. District Court for Northern California. CBS News reported that the plaintiffs seek to represent a nationwide class of customers.
The complaint says the companies coordinated efforts to slow artificial intelligence development, reducing the value customers receive from paid subscriptions. The companies had not responded to requests for comment by Saturday.
The case covers subscribers to ChatGPT, Claude, Grok and Gemini. It rests on antitrust laws, which bar competitors from making agreements that unlawfully restrain competition.
The allegations remain unproven, and the court has not ruled that any illegal agreement existed. Still, the filing raises a direct question: Who gets to control the pace of AI progress?
The lawsuit points to public cooperation among AI leaders on September 12. That day, Anthropic CEO Dario Amodei published an essay calling for industrywide cooperation to slow advances while companies pursued stronger safety measures.
The complaint says Sam Altman, Elon Musk and Google DeepMind’s Demis Hassabis agreed with Amodei’s proposal that same day. It describes his three-part plan as an effort aimed at “pacing the frontier.”
An earlier July 2026 statement acknowledged pressure on companies not to slow development on their own, the Associated Press reported. The plaintiffs argue that cooperation among leading rivals could produce slower progress than normal competition would.
The complaint draws a firm line between unilateral safety choices and joint action by competitors. The subscribers say they do not oppose companies independently emphasizing safety or asking the government to regulate the industry.
Their objection concerns private coordination. As the lawsuit puts it, “The antitrust laws do not permit competitors to decide among themselves that competition is too dangerous.”
Amodei wrote that federal mediation would help cross-company discussions, or that the government could at least enable them. He also suggested a “narrow waiver for certain kinds of safety conversations.”
Such a waiver could protect limited safety talks from antitrust concerns. But the available filings do not show that the government issued one or joined the discussions.
That gap matters. Companies may ask elected leaders for clear rules, but private firms cannot grant themselves legal permission to limit competition.
Amodei has framed the issue in stark terms. He warned that “swarms of rogue AI agents could take over the internet in as little as six months.”
In an interview with CBS senior business and technology correspondent Jo Ling Kent, Amodei also described his concern about the speed of development:
“I don't think I fully just appreciated what it would actually be like when the progress was as fast as it was,”
He later explained the stakes as he saw them: “If we build in the right way, I think the probability of something bad happening is very low.”
“If we build in the wrong way, the probability of something bad happening is very high,” Amodei said.
The plaintiffs’ theory is straightforward. Paid customers bought access to competing products, and competition should push each company to improve its service.
If rivals jointly slow their work, the complaint argues, subscribers receive less value than open competition would produce. No specific damages figure appears in the available account of the case.
Lead attorney Nick Rowley went further, warning against allowing powerful companies to set AI safety rules through private agreements. Breitbart quoted him as saying:
“AI will quickly spin out of human control and could kill us all if we allow AI safety and protocol... to be controlled by private self-serving agreements between the world's most powerful 'for profit' technology companies,”
Rowley’s warning attacks both sides of the arrangement. He contends that AI may carry grave risks, yet private coordination by profit-seeking rivals is not a sound answer.
The lawsuit now must establish that public calls for cooperation became an actual agreement restraining competition. Public expressions of support alone do not settle that issue.
Several basic questions also remain unanswered. The filing account does not identify the relief sought, list the plan’s three points or provide responses from the companies.
AI safety deserves serious rules made in public and under law. Consumers should not have to trust private competitors to draft those rules for themselves.
