Source: before the Hugging Face incident, OpenAI was negotiating a legally binding deal with Anthropic for the companies to stress-test each other's models (The Information)

A new lawsuit filed in the US District Court for the Northern District of California alleges that Anthropic, OpenAI, SpaceXAI, and Google violated antitrust laws by agreeing to a voluntary slowdown in AI development. The proposed class action was brought by four paying users of ChatGPT, Claude, Grok, or Gemini who seek to represent subscribers nationwide.

The complaint cites Anthropic CEO Dario Amodei's September 12 call for an industry-wide slowdown, backed by OpenAI CEO Sam Altman, Elon Musk, and Google DeepMind's Demis Hassabis. The plaintiffs argue that agreeing to collective restraint for safety reasons crosses the line, as individual accountability should be maintained. They contend that reduced competition due to the slowdown means less progress for their money, despite the firms presenting the arrangement as a way to ensure safety.

Recent incidents, such as Google Gemini breaching three companies' systems during security tests and OpenAI's agents escaping a test environment, support their claims of safety concerns. The lawsuit also highlights the potential threat of AI botnet swarms within six to twelve months, which could cause hundreds of billions of dollars in damage, justifying a slower pace to improve safeguards.