OpenAI Asks Regulators Whether an AI Slowdown Would Run Afoul of Antitrust Law

OpenAI has begun asking whether a coordinated slowdown in artificial intelligence development would be legal under U.S. antitrust law, according to a report from Wired. The question matters because the same companies that would need to slow down are the ones most exposed to price-fixing and collusion rules. If the law blocks cooperation, safety-minded pauses may be harder to arrange than the labs assumed.

The concern is not hypothetical. Antitrust statutes bar competitors from agreeing on output, prices or market division, and a joint decision to cap model training or delay releases looks like an agreement on output. AI executives have publicly floated slowdowns as a safety measure, but those discussions happen among a handful of firms that dominate frontier model development. That concentration is exactly what competition enforcers watch.

OpenAI's inquiry follows a year in which Washington has grown more aggressive on technology competition. The Federal Trade Commission and the Justice Department have pursued cases against large platforms, and the agencies have signaled interest in how AI markets form. Lina Khan's tenure at the FTC and Jonathan Kanter's antitrust division at DOJ produced guidance and investigations that treat concentrated control of compute and data as a competition problem.

There is a counterargument the labs can make. Antitrust law generally targets agreements that harm consumers through higher prices or reduced output, and a slowdown framed as safety research might be defended as a legitimate, pro-competitive action. Legal experts have debated whether a short, transparent pause could survive review if it were limited in scope and paired with public disclosure. No court has tested that theory for AI.

The practical stakes are high because compute and talent are finite. If one lab slows down while rivals keep training larger models, the cautious firm risks losing ground, a dynamic that makes voluntary restraint unstable without coordination. That is why the industry keeps circling the same question: can competitors legally agree to move slower, or does the law force them to keep racing.

What to watch next is whether OpenAI's questions turn into formal requests for guidance or public statements from the FTC and DOJ. Any written position from the agencies would shape how labs structure safety commitments, and it could determine whether an industry-wide pause is ever more than a talking point. For now, the answer remains unresolved, and the labs are asking before they act.