FTC chair is 'deeply suspicious' of the labs' antitrust ask
Washington got two answers on AI policy within a day: the administration's competition chief pushed back on the industry's safety-for-antitrust trade, and OpenAI told Congress it will take a mandate it once argued against.
FTC Chairman Andrew Ferguson said "everyone should be deeply suspicious" of AI companies that come to Washington asking for new regulations and an antitrust exemption at the same time. Speaking at Georgetown University on Tuesday, Ferguson said that combination is a request for "barriers to entry that will insulate their incumbency from challenge," adding that "all of my alarm bells go off." He said he was giving his personal opinion and did not name a company, but the target was unmistakable: Anthropic CEO Dario Amodei's proposal, made last week, to let rival labs coordinate on a slower frontier while models capable of shutting down parts of the internet get tested more aggressively. Ferguson's comments are the first public indication of how the Trump administration reads that pitch.
The tell is structural. The exemption Amodei wants is not something a lab can grant itself — coordination between competitors is the classic antitrust fact pattern, so the labs need either the FTC and DOJ to look away or Congress to write permission into law. Ferguson just publicly questioned the premise in front of a policy audience, and he pointed out that President Donald Trump sets federal AI policy, which means the ask now depends on a White House that spent Monday calling safety fears a hoax. The labs are asking two different branches for two different things: less scrutiny of how they coordinate, and a federal framework that binds everyone else. We covered the first version of this ask — OpenAI asked Congress if an industry-wide AI slowdown is even legal.
OpenAI has told a sponsor of the FRONTIER Act that it can support the bill's requirement that frontier developers embed outside evaluators — the first time the company has backed a specific federal third-party safety mandate. Politico reported Tuesday that OpenAI's lobbyist made the commitment in a Monday meeting on Capitol Hill, telling the sponsor's office it supports the bill's independent-verification provision and that "we wanted to be really clear about that." A Trahan staffer called it "a really strong signal that the leaders of these labs jumped on board with Dario's embedded evaluators idea," while cautioning that the next few weeks will show whether the commitment survives contact with actual bill text.
The bill in question is bipartisan, not new, and bigger than the ask. The FRONTIER Act, introduced in July by Representatives Jay Obernolte and Lori Trahan with four co-sponsors split evenly between the parties, sets tiered duties on the largest developers — model cards, risk-management frameworks, independent audits, incident reporting and ongoing assessments — and would preempt state "substantive obligations" in the same stroke. That preemption clause is the real prize and the real fight: it is why states and some Democrats have resisted, and why an industry weary of a fifty-state patchwork has kept talking. OpenAI's endorsement is narrow — one provision, not the bill — but it moves the lab from pledging voluntary evaluators to asking Congress to make evaluators mandatory, which is a different category of commitment. We covered the voluntary half of that shift — OpenAI wants national AI rules — and CAISI's evaluator joins its board.
What to watch: Whether House Energy and Commerce marks up the FRONTIER Act before the fall recess — House Democratic leader Hakeem Jeffries has already called for Congress to stay in session — and whether Anthropic answers Ferguson directly or lets the White House answer for it.
Should a lab that asks for a federal safety mandate get an antitrust pass for the coordination that mandate requires? Tell us in the comments.
