A new lawsuit has accused Anthropic, OpenAI, xAI and Google of making an illegal agreement to slow the development of their artificial intelligence systems.
The lawsuit, filed Friday in the US District Court for the Northern District of California, alleges that the companies violated antitrust laws by coordinating efforts to slow AI development.
The plaintiffs argue that such coordination could reduce competition and limit the value consumers receive from paid AI subscriptions.
The lawsuit was filed on behalf of four named plaintiffs who pay for subscriptions to ChatGPT, Claude, Grok or Gemini. The plaintiffs are seeking to represent a nationwide class of other paid subscribers to the services.
What the lawsuit alleges
The lawsuit points to public comments made by leaders of the four companies after Anthropic CEO Dario Amodei called for the AI industry to slow the pace of development.
According to the complaint, the coordination largely took shape on Sept. 12, when Amodei published an essay calling for industrywide cooperation to slow advances in AI capabilities and give companies more time to strengthen safety measures.
OpenAI CEO Sam Altman, xAI CEO Elon Musk and Google DeepMind co-founder and chair Demis Hassabis publicly responded to Amodei’s proposals the same day, with Altman and Musk expressing agreement.
The lawsuit also alleges that discussions about slowing AI development began months earlier. It cites a July 2026 statement signed by senior employees from several leading AI companies that acknowledged the “intense competitive pressure not to unilaterally slow” development.
The statement called for government support for a global effort to slow the development of increasingly capable AI systems.
The plaintiffs argue that an agreement between major AI competitors to make technological progress slower than it would otherwise be has an anticompetitive effect on consumers.
What Amodei said
Amodei has called for AI companies to slow the pace at which they improve the capabilities of their models while continuing technical progress.
“We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain,” Amodei wrote in his essay.
His proposal included three main steps: independent evaluators with access to companies’ safety practices, coordination among leading AI companies on safety standards and limits on unchecked development, and international cooperation on AI risks.
Amodei also warned about the speed at which AI systems could become capable of causing large-scale harm.
“Given the accelerating rate of AI capability development, it’s my worry that in 6-12 months such a swarm could be capable of taking over the entire internet potentially causing hundreds of billions of dollars in damage,” he wrote.
Amodei said he was not calling for a halt to AI model training or technical progress. He argued instead that companies should take adequate time to align and safeguard their models and allow independent evaluators to verify those measures.
Altman and Musk backed the proposal
Altman said OpenAI agreed with the need to pace frontier AI development. “I agree with Dario that we need to pace the frontier,” Altman said in a post on X. He also said independent evaluators having employee-like access was a good idea and that OpenAI would do the same.
Musk also backed Amodei’s proposal, writing: “Dario is right.” Musk has previously warned about the risks posed by increasingly capable AI systems.
Plaintiffs say companies can act independently
The plaintiffs said they do not object to the companies independently choosing to slow their own development because of safety concerns.
Their objection is to the companies reaching an agreement with one another to limit competition, according to the complaint. The lawsuit argues that the companies could instead ask Congress, the White House or other government agencies to establish AI safety rules or seek an antitrust exemption for specific safety-related discussions.
In his original essay, Amodei acknowledged the potential antitrust concerns. He said the US government could mediate discussions between AI companies or “at least enable” them by issuing a narrow waiver for certain types of safety conversations.
The lawsuit comes as AI companies face growing scrutiny over the risks associated with increasingly capable models, including concerns about cybersecurity, misuse and whether AI systems could eventually operate beyond effective human control.
Anthropic recently disclosed cases involving its Claude models being used or tested in activities involving cybersecurity, surveillance, fraud and weapons development.
The company has also reported instances in which its models were able to hack external systems during cybersecurity tests.
Debate over government involvement
Altman has said OpenAI supports a “federal framework that sets consistent safety requirements”. He also said the industry did not need to wait for an antitrust exemption or new legislation before beginning safety work.
The question of whether AI companies should be allowed to coordinate on safety has also drawn opposition from some US lawmakers.
Sen. Josh Hawley, a Republican from Missouri, said during a recent Senate hearing that he would not support giving major technology companies an antitrust exemption to collaborate, arguing that such an arrangement could allow them to stifle competition.
The lawsuit now puts the companies’ public calls for cooperation under an antitrust lens, with the plaintiffs arguing that safety coordination should not come at the expense of competition.