US FTC chief warns against antitrust carve‑outs for AI safety coordination

Ferguson’s remarks are a clear shot across the bow of Anthropic and any other frontier lab flirting with the idea of an explicit slowdown pact. His core argument is that when the same firms asking Washington for strict AI rules also request antitrust immunity to coordinate, regulators should assume they are trying to lock in incumbency under the guise of safety. That framing will shape how any future “AI OPEC”‑style proposals are received on Capitol Hill.

In AGI terms, this makes a synchronized, regulation‑backed pause less likely in the near term. If labs know that formal coordination on capability throttling could trigger antitrust action, they will instead favor softer, voluntary pacing and safety commitments that stop short of hard caps. That may keep the competitive race hot even as public rhetoric shifts toward caution.

At the same time, Ferguson’s line that President Trump will ultimately set AI policy underscores how politicized this space has become. If executive‑branch signals tilt toward maximizing US advantage against China, agencies like the FTC will face pressure to treat safety‑driven coordination as collusion while turning a more permissive eye to aggressive expansion. The net result could be that safety cooperation happens mostly through informal networks and shared tools, while capital and compute keep pouring into the frontier.