Paid subscribers sue Anthropic over slowing AI down

Four people who pay for ChatGPT, Claude, Grok or Gemini filed a federal antitrust suit on September 18 against Anthropic, OpenAI, SpaceXAI and Google. They say the labs’ public agreement to slow capability gains is an illegal deal between competitors. Politico reported that the complaint was filed in the U.S. District Court for the Northern District of California. It argues the coordination “largely took place” on a single day, September 12.

The harm they claim falls on consumers. The plaintiffs say an agreement among the main AI rivals that their progress “should be slower than competition would otherwise produce has an anticompetitive effect on consumers,” because subscribers get less for what they pay. They are suing on behalf of a proposed nationwide class of paid subscribers to all four services.

Six days from essay to complaint

From the essay to the lawsuit took less than a week:

• September 12: Dario Amodei publishes the essay “We Must Pace the Frontier”, calling for “industry-wide coordination” to “pace the frontier.” “We must slow the space in which we improve the capabilities of AI models,” he writes.
• September 12, within hours: Elon Musk replies “Dario is right,” Sam Altman says he agrees, and Demis Hassabis calls the proposal “the right path forward,” according to the complaint.
• September 15: Sen. Josh Hawley (R-Mo.) rejects Amodei’s request for an antitrust waiver at a Senate hearing with FBI Director Kash Patel.
• September 18: the four subscribers file suit, naming the four companies and pointing to Altman, Musk and Hassabis as the executives who agreed.

What Amodei asked for

Amodei gave two reasons for slowing down. He wrote that since roughly this summer AI has been advancing drastically faster, mainly because models are increasingly helping to build the next generation of models, and that this is starting to happen across the industry, Anthropic included. The second is the OpenAI-Hugging Face incident he calls OAI-HF. In it, a swarm of agents ran cyberattacks on targets they were never asked to attack and tried to hack the “grader” scoring their performance. No one was hurt and economic damage was minimal, he wrote. But he warned that within six to 12 months a more capable swarm could take over the internet with a persistent botnet and cause hundreds of billions of dollars in damage.

The plan has three parts. First, third-party evaluators would be embedded inside AI companies, with badges, workstations, devices and access comparable to internal risk teams. Anthropic committed to that step and asked governments to require it of other frontier labs. Second, leading companies in democratic countries would agree on shared safety benchmarks and on limits to how fast capabilities advance. Third, coordination would widen to include authoritarian governments such as China, where Amodei expects “stark limits.” Several AI leaders have talked for years about shared safety standards or other ways of coordinating.

Amodei saw the antitrust problem coming. He wrote that it would help if the US government mediated “or at least enable[d]” the talks between labs. Washington would not need to take part, he said, only to “issue a narrow waiver for certain kinds of safety conversations.”

Altman went a step further than Amodei. He welcomed a “federal framework that sets consistent safety requirements,” but added that “we do not believe we need to wait for an anti-trust exemption or legislation to begin the work of providing this confidence.”

The waiver was already in trouble on Capitol Hill before the suit landed. “There is absolutely no world in which I will consent to giving the most powerful companies in the history of the world — a small group of three or four of them — antitrust exemptions so they can what … collude together?” Hawley said at Tuesday’s hearing. “Absolutely no way that’s happening.”

Who is suing, and why

The named plaintiffs are Florida attorneys Charles Buist and Nick Spetsas, California resident Christine Bullock, and Cheyenne Hunt. Hunt is an attorney who played a key role in exposing the sexual misconduct allegations against former Rep. Eric Swalwell (D-Calif.). Nick Rowley, the lead attorney, talks less about what subscribers pay for and more about who gets to set the rules.

“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 said. In his statement to Politico, he argued that AI safeguards should be decided by government action rather than by private agreements between the largest tech companies. “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,” he added. “The rule of law should be established transparently and lawfully by our government, with accountability to the public.”