|
Getting your Trinity Audio player ready...
|
Four of the world’s most powerful AI companies are now defendants in a federal antitrust lawsuit that accuses them of illegally agreeing to slow down their own technology, an unusual legal theory that turns the companies’ own public statements about AI safety into the core evidence against them. The complaint, filed Friday in the U.S. District Court for the Northern District of California, names Anthropic, OpenAI, SpaceXAI and Google as defendants and alleges the companies coordinated to restrict competition rather than each independently deciding to pursue caution on its own.
The lawsuit centers on a sequence of public statements made on September 12. That day, Anthropic CEO Dario Amodei published an essay titled “We Must Pace the Frontier,” arguing that AI companies should slow the rate at which they improve model capabilities and calling for industry-wide coordination on safety pacing. According to CBS News, Amodei’s essay warned that swarms of rogue AI agents could potentially take over significant portions of the internet within as little as six months absent more careful development practices. Within hours, according to the complaint, OpenAI CEO Sam Altman, SpaceXAI founder Elon Musk and Google DeepMind co-founder Demis Hassabis each publicly responded in agreement, with Musk writing that Dario is right and Hassabis calling the proposal the right path forward, according to reporting from The Hill.
The plaintiffs, four paid subscribers to the companies’ AI products represented by attorney Nick Rowley, argue that this public back-and-forth amounted to an illegal restraint of trade under the Sherman Act rather than a coincidental alignment of independent safety views. Their central argument is not that slowing AI development is wrong in principle. According to Tom’s Hardware, the plaintiffs explicitly acknowledge the need for the industry to decelerate for safety reasons, but argue that Amodei’s call for coordinated action represents a shortcut that substitutes collective restraint for individual accountability, allowing competitors to avoid the market pressure that would otherwise force each company to weigh safety against competitive risk on its own. The complaint further argues that reduced competition on the pace of innovation could diminish the value subscribers receive from paid AI products, tying the antitrust claim directly to consumer harm.
Notably, the lawsuit alleges the coordination did not begin on September 12 itself but had been developing for months beforehand, suggesting the public essay and immediate responses represented the visible culmination of private discussions rather than a spontaneous, independent convergence of views among competing CEOs. That claim, if substantiated through the discovery process, would significantly strengthen the antitrust theory, since coincidental public agreement among rivals is generally far harder to prosecute under the Sherman Act than evidence of prior private coordination.
Rowley’s public statements about the case have leaned heavily into the same existential framing that Amodei himself used to justify the original slowdown proposal, just aimed in the opposite direction. In a statement reported by The Hill, Rowley argued that humanity deserves ironclad safeguards against extinction-level threats and described the current situation as the biggest risk to mankind in history, while separately telling reporters that allowing AI safety protocols to be controlled by private, self-serving agreements between powerful for-profit companies risked exactly the kind of uncontrolled outcome those companies claim to be trying to prevent.
Anthropic had anticipated this exact legal vulnerability. According to multiple outlets covering the case, Amodei’s original essay explicitly acknowledged the antitrust risk inherent in calling for industry coordination, expressing hope that government regulators might grant some form of exemption or establish a legal framework allowing safety-focused cooperation among competitors without triggering standard antitrust liability. Altman responded publicly along similar lines at the time, stating that OpenAI would welcome a federal framework setting consistent safety requirements across the industry, while adding that the companies did not believe they needed to wait for a formal antitrust exemption or new legislation before beginning that work.
That request for regulatory accommodation has not been granted. President Donald Trump rejected calls for a coordinated regulatory framework around AI safety pacing over the weekend, according to Ground News, leaving the four companies to face the antitrust claim without the kind of government cover Amodei had hoped for when he first raised the antitrust risk publicly. Representatives for Anthropic, OpenAI, Google and SpaceXAI did not immediately respond to requests for comment following the lawsuit’s filing, according to the Associated Press.
The case arrives at a moment when public anxiety about AI safety and public suspicion of Big Tech coordination are both running unusually high simultaneously, a combination that puts the defendant companies in a genuinely difficult position regardless of the lawsuit’s ultimate legal merits. Companies that publicly commit to safety-driven caution risk exactly this kind of antitrust exposure when rivals visibly agree with them, while companies that decline to coordinate publicly on safety risk separate criticism for prioritizing competitive advantage over stated safety concerns. That tension is unlikely to resolve quickly, and how courts eventually treat public, near-simultaneous statements of agreement among competing AI CEOs as either innocent parallel conduct or actionable coordination could shape how cautiously AI companies discuss safety measures publicly going forward, well beyond the outcome of this specific case.
For now, the lawsuit remains in its early stages, and none of the four companies has filed a formal response in court. Given the significant financial and reputational stakes involved for all four defendants, and the novel legal theory at the center of the complaint, the case is likely to draw sustained attention from antitrust scholars, AI policy observers and the companies’ own investors as it moves through the federal court system. Continuing coverage of how legal and regulatory challenges are shaping the AI industry is available on Business Tech. Amodei’s original essay proposing the safety pacing framework is available through Anthropic’s official blog, and further reporting on the lawsuit can be found through PBS NewsHour’s coverage of the case.
