- Four subscribers have sued Anthropic, OpenAI, SpaceXAI, and Google over an alleged pact to slow AI improvements.
- The proposed class action claims the deal violates the Sherman Act by restricting output.
- The suit follows Anthropic CEO Dario Amodei's call to slow AI development, which rival executives endorsed.
- Plaintiffs argue AI safety rules belong with government, not private corporate agreements.
- Sen. Josh Hawley has rejected granting AI companies antitrust exemptions to coordinate.
A proposed class-action lawsuit filed Sept. 18 in federal court in Northern California accuses Anthropic, OpenAI, SpaceXAI, and Google of conspiring to deliberately slow improvements to their artificial intelligence models, allegedly violating the Sherman Act. Four subscribers who paid for subscriptions to Claude, ChatGPT, Grok, or Gemini argue they were promised continuous upgrades and are now receiving lesser value due to what they call an illegal agreement among competitors to restrict output.
The legal action follows Anthropic CEO Dario Amodei's public call for industry-wide coordination to "pace the frontier" of AI development, a proposal that drew quick endorsements from OpenAI's Sam Altman, SpaceXAI's Elon Musk, and Google DeepMind co-founder Demis Hassabis.
Subscribers claim coordinated slowdown cheats paying customers
According to the lawsuit, customers pay for these services to get each company's strongest models along with a steady stream of upgrades. "An agreement that slows improvement lowers the quality of what subscribers receive for the price they pay," the complaint states.
Plaintiffs allege the pact began taking shape in July, when leading AI labs put their names to a statement conceding they faced strong competitive pressure not to slow down on their own. According to the plaintiffs, the effort came into the open with Amodei's Sept. 12 blog post calling for slower capability improvements. Musk endorsed it within hours, and Altman and Hassabis also voiced support.
Legal team says safety concerns don't justify private deals
Attorney Nick Rowley, one of the lawyers representing the plaintiffs, argues that decisions about AI safety belong in the hands of government, not corporate boardrooms. In a statement to Politico, he compared the threat to nuclear warfare and called AI the greatest danger humanity has ever faced.
Rowley warned that "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."
Senator rejects calls for antitrust exemption on safety grounds
The controversy has already reached Capitol Hill. Amodei had suggested the government could grant AI companies a limited antitrust waiver so they could discuss safety standards together. Sen. Josh Hawley (R-Mo.) dismissed the idea at a Sept. 15 Senate hearing with FBI Director Kash Patel: "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?"
The lawsuit argues each company is free to decide its own safety limits and safeguards without striking a deal with competitors. While plaintiffs acknowledge the importance of safety, they call Amodei's proposal a "shortcut" that trades individual accountability for collective restraint and ultimately harms consumers.
Private pact raises questions over who controls AI's pace
The four companies did not respond to requests for comment from The Epoch Times. The lawsuit represents a clash between legitimate safety concerns and the foundational principles of competitive markets. The proposed slowdown also runs counter to the White House's agenda: President Donald Trump announced Sept. 19 that he would form an "AI Force" to boost the industry's growth.
As the legal battle unfolds, the fundamental question remains: Should a handful of powerful companies be permitted to collectively decide the pace of technological advancement, or should market forces and government regulation determine how fast AI evolves? The executives now urging restraint are the same ones who built these systems and sell access to them. If AI is as dangerous as they claim, those decisions belong with officials who answer to voters, not with a few firms that stand to profit either way.
Sources for this article include:
Please contact us for more information.
