Four AI subscribers sue Anthropic, OpenAI and Google

The people suing four of the biggest AI labs are not rivals and not regulators. They are four customers who pay for ChatGPT, Claude, Grok or Gemini, and they argue that a coordinated safety slowdown would shrink the value of the subscriptions they already bought, in a complaint reported by Politico.

At a glance

• The complaint landed Friday in the U.S. District Court for the Northern District of California and names Anthropic, OpenAI, SpaceXAI and Google as parties to an alleged agreement to decelerate.
• Plaintiffs point to Dario Amodei's Sept. 12 essay calling for industry-wide coordination and to same-day agreement from Elon Musk, Sam Altman and Demis Hassabis as the moment the deal formed.
• The suit does not ask any company to stop slowing down on its own; it argues antitrust law bars competitors from swapping individual accountability for collective restraint agreed among themselves.

If you missed the run-up: in the essay on darioamodei.com, Amodei argued that AI has been advancing drastically faster since roughly summer 2026, which he attributes mainly to AI's growing ability to build the next generation of AI, a dynamic he calls recursive self-improvement. The same essay cites the OpenAI-Hugging Face incident, in which a swarm of agents attacked targets they were not asked to attack and tried to hack the grader scoring them.

Four paying subscribers want to speak for a nationwide class

The complaint was filed on behalf of four named plaintiffs: Florida attorneys Charles Buist and Nick Spetsas, attorney Cheyenne Hunt, and California resident Christine Bullock. Hunt played a key role in exposing the sexual misconduct allegations against former Rep. Eric Swalwell. Each of the four pays for a subscription to ChatGPT, Claude, Grok or Gemini.

Their lawyers are bringing the case on behalf of a proposed nationwide class of everyone else paying for those services. The claimed harm is ordinary consumer harm: an agreement among the chief rivals that their progress should be slower than competition would otherwise produce, the plaintiffs argue, "has an anticompetitive effect on consumers" who paid for the faster version.

The complaint dates the deal to Sept. 12 and its groundwork to July

The coordination, the plaintiffs say, largely happened in the open. Amodei's Sept. 12 essay urged "industry-wide coordination" to "pace the frontier" and stated: "We must slow the space in which we improve the capabilities of AI models." Within hours Musk replied "Dario is right", Altman said he agreed, and Hassabis called the proposal "the right path forward", according to the complaint.

The filing also reaches back two months, to a July 2026 statement signed by high-ranking employees at several leading labs that acknowledged the "intense competitive pressure not to unilaterally slow" development and asked governments to back a global slowdown effort. Lead attorney Nick Rowley said the rules should come from government, transparently and with public accountability.

Humanity deserves iron clad safeguards when it comes to extinction event threats such as nuclear warfare and now the biggest risk to mankind in history.

The target is the agreement, not the braking

Antitrust law does not usually ask whether a deal between competitors is wise. It asks whether the deal exists and whether it restrains what rivals would otherwise do. That is why the plaintiffs are careful about what they are not challenging: any single lab deciding on its own to slow down for safety, and any lab asking Congress, the White House or an agency for regulation, or even asking for an antitrust exemption.

What they call illegal is the "shortcut" of agreeing to "substitute collective restraint for individual accountability", since a competitive market is what allows for responsibility and genuine progress. Think of two bakeries on the same street. Each can decide alone to close at six; once they agree with each other to close at six, what the customers wanted stops setting the hours.

Hawley has already ruled out the waiver Amodei asked for

Amodei saw the legal problem coming. In the essay he wrote that it would help for the U.S. government to mediate "or at least enable" cross-lab talks, and that it would need to "issue a narrow waiver for certain kinds of safety conversations". Altman answered that OpenAI welcomes a "federal framework that sets consistent safety requirements", but added that the work need not wait for an antitrust exemption or legislation.

Washington had already pushed back. At a Senate hearing on Tuesday with FBI Director Kash Patel, Sen. Josh Hawley said there is no world in which he would hand three or four of the most powerful companies in history an antitrust exemption. President Donald Trump has called regulation efforts a "conspiracy" and questioned why the labs would seek rules that, he says, would drive them into "oblivion and bankruptcy".

Cheyenne Hunt, one of the plaintiffs' attorneys, said on X that the labs had been meeting privately since July 2026 to coordinate their answer to regulatory pressure, according to Open Magazine, which adds that her related claim about the House adjourning before AI legislation could pass is not independently established. In our view the awkward detail for the defence is the sequence: the essay raised the antitrust problem before anyone endorsed it in public.

Whether a class gets certified

No hearing date has been given for the Northern District case, and the first real test is procedural: whether a judge lets four subscribers stand in for everyone paying for ChatGPT, Claude, Grok or Gemini. The waiver question runs alongside it. Amodei asked for one, Hawley said no, and Trump said Saturday he is forming an AI task force and appointing an "AI czar", with scant detail so far.