OpenAI asks Congress: is an AI slowdown legal?

OpenAI's chief scientist, Jakub Pachocki, published a blog post last weekend arguing that the best path forward for the AI research world includes "coordinating to slow down future development," and that in the short term he expects "voluntary slowdowns to become commonplace until shared safety bars are established." OpenAI's push to Congress for guidance is a separate, earlier track: the company has spent recent weeks asking lawmakers whether orchestrating an industry-wide slowdown on frontier AI development would be legal, according to people close to the company.

The legal obstacle is antitrust. Nicholas Felstead, assistant director of the Australian Competition and Consumer Commission and a former AI policy fellow at the Center for Law & AI Risk, argued in a March article that a coordinated pause in AI development may amount to companies restricting output, potentially violating the Sherman Antitrust Act. He wrote that it would depend "entirely on the precise details of any agreement," but that "even if most safety collaborations would ultimately survive antitrust scrutiny, legal uncertainty can act as a powerful deterrent." Some lawmakers appear to be listening: in July, a bipartisan, bicameral group introduced the "Collaboration on Adversarial Threats and Security Risks Act," which would explicitly permit AI labs to coordinate on security and safety work without the risk of violating antitrust statutes. The House version was referred to the Judiciary Committee but has yet to be taken up, and Caleb Knapp, director of government affairs at the AI Policy Network, which endorsed the bill, says enacting anything into law may have to wait until after the upcoming midterm elections.

Not everyone accepts the antitrust framing. John Schulman, an OpenAI cofounder who is now chief scientist at the rival AI lab Thinking Machines, wrote on X earlier this week that "first step is for industry leaders OpenAI and Anthropic to stop feuding and work on a pacing proposal together," adding "They'll cite antitrust, but that's fake—antitrust prohibits certain agreements, but not from jointly developing a proposal." The article notes that some AI leaders see antitrust worries as convenient cover, with other reasons for reluctance including fierce competition for the nascent frontier-model market, some executives' shared view with the Trump administration that staying ahead of China is crucial for national security, and differing opinions among developers about how best to develop safe AI. The backdrop has sharpened: this week former Anthropic and OpenAI researcher Jacob Coxon issued a stark public warning that AI developers were putting humanity at risk, and recent security incidents, including OpenAI's agents hacking Hugging Face, have highlighted safeguards failing to keep pace with rising model capabilities. Whether any coordinated slowdown becomes legally workable appears to hinge on whether Congress provides explicit antitrust cover, and on whether labs can agree on a shared safety approach in the meantime.