OpenAI has reached out to Congress for clarity on whether orchestrating a slowdown in AI development across the industry would violate antitrust laws, sources tell WIRED. Substantive coordination on safety between labs could risk contravening these laws, posing a barrier to major tech firms joining the initiative.

The best path forward, argues OpenAI’s chief scientist, Jakub Pachocki, is a “voluntary slowdown” until safety standards are established. However, legal scholars warn that such efforts could conflict with US antitrust statutes. Nicholas Felstead, a former AI policy fellow, suggests that such agreements may be seen as companies restricting output, potentially violating the Sherman Antitrust Act.

Early signs suggest Congress might address this issue. A bipartisan bill, titled “Collaboration on Adversarial Threats and Security Risks Act,” was introduced in July and aims to allow AI labs to coordinate on safety and security without fear of antitrust violations. However, enacting anything into law may have to wait until after the upcoming midterm elections.

Despite these concerns, some industry leaders argue that antitrust worries are merely a convenient cover. John Schulman, chief scientist at rival AI lab Thinking Machines, suggests that industry leaders should stop feuding and work on a pacing proposal together. He believes the real reasons for reluctance to collaborate are more about business competition and differing opinions on AI safety.