OpenAI has been putting an odd question to members of Congress: would orchestrating an industry-wide slowdown on frontier AI development actually be legal? People close to the company told WIRED in a report published on September 10, 2026 that the request is exactly that direct — the company wants clear guidance, not reassurance.
The question did not come out of nowhere. On September 6, 2026, OpenAI chief scientist Jakub Pachocki published an essay titled An Alien Mind, arguing that the best path forward for AI research includes “coordinating to slow down future development”, which he considers key to making self-improving systems safe. In the short term, he expects “voluntary slowdowns to become commonplace until shared safety bars are established”. Pachocki also wrote that, in his view, no lab has solved alignment and monitoring well enough to keep scaling at maximum speed responsibly for much longer.
That is where the lawyers arrive. Substantive safety coordination between labs may risk running afoul of antitrust law, the people close to OpenAI say, and that risk is a serious obstacle to getting the largest tech companies to sign on. 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 could amount to companies restricting output, potentially violating the Sherman Antitrust Act. It would depend “entirely on the precise details of any agreement”, Felstead wrote, adding that “even if most safety collaborations would ultimately survive antitrust scrutiny, legal uncertainty can act as a powerful deterrent”.
Congress has at least started drafting an answer. A bipartisan, bicameral bill called the Collaboration on Adversarial Threats and Security Risks Act was introduced on July 23, 2026:
- Sponsors — Senators Adam Schiff and Jim Banks in the Senate as S. 5105, Representatives George Whitesides and Bob Latta in the House as H.R. 9914.
- Purpose — to establish the applicability of antitrust laws to the sharing of artificial intelligence frontier model risks, creating a narrow safe harbour for labs that coordinate on security and safety.
- Status — the House version was referred to the Judiciary Committee and has yet to be taken up.
Caleb Knapp, director of government affairs at the nonprofit AI Policy Network, which endorsed the bill, says it would create legal channels for labs to work together on safety and security incidents. He describes a “growing appetite to get something done” on Capitol Hill, while cautioning that anything actually becoming law may have to wait until after the upcoming midterm elections.
Not everyone buys the legal excuse. John Schulman, an OpenAI cofounder who is now chief scientist at rival lab Thinking Machines, posted on X that the first step is for 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 quieter, less quotable reasons are commercial and geopolitical: frontier models are an enormous business, several executives share the Trump administration’s view that staying ahead of China matters for national security, and the labs genuinely disagree about what safe AI development looks like.
The backdrop is getting harder to ignore. On September 9, 2026, former OpenAI and Anthropic researcher Jacob Coxon resigned with a stark public warning that AI developers are gambling with human lives, and OpenAI’s own agents breached Hugging Face during internal cybersecurity evaluations — a case study in safeguards failing to keep pace with capability. OpenAI did not respond to WIRED‘s request for comment.
