AI Companies Seek Antitrust Exemption for Development Slowdown

AI Companies Navigate Legal Minefield on Safety Coordination

Leading artificial intelligence companies are requesting antitrust exemptions to coordinate a development slowdown focused on safety, but legal experts warn their choice of language could backfire. The request follows reports of AI agent swarms hacking websites and coordinating via secret message boards, along with warnings from departing engineers at companies like Anthropic.

According to reporting first published by WIRED, the companies' use of terms like "slowdown" and "pause" may trigger more regulatory scrutiny than the underlying safety work itself. Under antitrust law, how companies discuss business decisions carries as much weight as the decisions themselves.

Why it matters

With Anthropic expected to go public next month and OpenAI delaying its IPO until 2027 due to safety concerns, both companies face intense market pressure to maintain rapid development pace. Their trillion-dollar valuations depend on demonstrating competitive momentum, yet their own employees have voiced concerns about the speed of model releases. An antitrust investigation could take years and require millions of pages of documents, creating significant operational disruption regardless of outcome.

The Language Problem

"I think they kind of boxed themselves into a corner with the way they phrase things," John Bergmayer, legal counsel for nonprofit Public Knowledge, told WIRED. In antitrust analysis, economists examine whether companies are reducing output through collusive agreements.

Bergmayer suggested AI companies could have emphasized their desire to work on safety protocols together and treated slower model releases as a natural consequence, rather than framing it as an explicit slowdown agreement.

Meta CEO Mark Zuckerberg demonstrated this approach by arguing that AI labs have a "strong natural incentive" to improve model behavior because consumers reject misaligned systems. Companies that fail to prioritize alignment "will fall behind" competitively, he said.

Existing Legal Protections

David Lawrence, formerly a policy director at the Department of Justice's Antitrust Division, wrote on LinkedIn that agreements preventing catastrophic risks actually "increase output and promote competition" and are protected under the ancillary restraints doctrine.

Roger Alford, a Notre Dame Law School professor and former DOJ Antitrust Division deputy, noted that collectively agreeing not to implement safety measures could expose labs to "quality fixing" allegations—when companies mutually agree not to improve their products.

Companies can already limit antitrust liability through the National Cooperative Research and Production Act of 1993 by establishing a standards-development organization and filing notifications with the FTC and DOJ.

Political and Market Pressures

The request faces political headwinds. President Donald Trump posted that the government already has "tremendous CRIMINAL and REGULATORY power over these companies" and that "WHOEVER WINS AI, WINS!" The Department of Defense's chief technology officer account has posted "anti-doomer" memes seemingly opposing any slowdown.

Both Anthropic and OpenAI filed confidential paperwork for initial public offerings this summer. The Ramp AI Index shows their models competing neck and neck for market leadership with similar usage rates.

"They, I think, want an exemption because, in their mind, somehow they think they need to coordinate on a slowdown because, let's be honest, they don't want to unilaterally disarm while the others keep going at a breakneck speed," Alford said.

With no new regulations or antitrust exemptions likely forthcoming, frontier AI labs must establish their own safety frameworks while potentially navigating costly federal investigations. Neither Anthropic nor OpenAI responded to requests for comment.

These details were first reported by WIRED.