Skip to content
NewsRegulation

Trump rejects AI pacing call; Sacks tells labs to slow voluntarily but without antitrust cover

· by Pondero Newsdesk

The short version

President Trump dismissed frontier AI slowdown calls on September 13, and former White House AI czar David Sacks told Anthropic and OpenAI they can pace voluntarily but should not expect an antitrust waiver or a government approval process in return.

Trump rejects AI pacing call; Sacks tells labs to slow voluntarily but without antitrust cover

Two responses arrived on September 13 that drew a hard perimeter around the AI pacing coalition Anthropic CEO Dario Amodei launched one day earlier. President Trump rejected the call outright from Ireland. Former White House AI czar David Sacks rejected the mechanism attached to it: no antitrust carve-out, no government approval process, no official backing of any kind.

What happened

Speaking to reporters at his Doonbeg golf resort in Ireland, Trump dismissed AI safety warnings as "exaggerated" and attributed them to unnamed "negative forces." His full comment, per NPR: "We're leading China in AI, we're the most sophisticated country in the world, and frankly, I want to keep it that way because whoever wins AI, wins. And we can put guardrails, we can do this and that. But I think you have a lot of negative forces that are bringing it up that shouldn't be bringing it up." House Speaker Mike Johnson echoed the position on CNN's State of the Union that same day, saying any Congressional moratorium on AI would mean "China will overlap us."

Sacks, now co-chair of the President's Council of Advisors on Science and Technology, posted a separate response on X. He told Altman and Amodei to "go ahead" with voluntary slowing but drew a firm line on the conditions. "Stop pretending you need anyone else's permission," he wrote. He argued that the labs already face "massive product-liability exposure" if their models enable a serious cyberattack and that this pressure is sufficient discipline without any formal approval regime. Per TNW's reporting on the same post, Sacks wrote that labs asking for antitrust protection as the price of restraint were effectively asking to "form a cartel" and that framing such a trade as purely altruistic would look like "blackmail."

Sacks also challenged the independence of METR, the nonprofit evaluator Anthropic relies on for safety audits, per TNW, saying it is intertwined with the company's investors and staff.

This exchange followed Amodei's September 12 essay "We Must Pace the Frontier," in which he called for labs to stretch timelines between capability leaps and embed third-party evaluators to report safety incidents. Sam Altman and Elon Musk publicly backed the essay on social media. Altman separately told Fortune the OpenAI IPO is "ill-timed" given safety concerns and will not happen before 2027, per Fortune. Bloomberg's September 14 coverage also documented investor pressure from Wall Street, where concern over the AI valuation premium has added a financial dimension to the headwinds facing any organized pacing effort.

Why it matters

Two legal and political paths toward structured AI pacing are now explicitly closed. The Trump administration will not grant antitrust protection for competing labs to coordinate development timelines, and it will not stand up a government approval process with authority over frontier releases. Sacks' framing redirects the enforcement question to product liability courts rather than a regulatory body.

For AI tool operators, the practical consequence is concrete: any "safety pledge" from a frontier lab is a unilateral commitment absorbing competitive cost alone. There is no coordination mechanism available. A lab that slows down gives ground to labs that do not, and no government authority will insulate it from that trade-off.

The Sacks critique of METR is worth tracking separately. If the independence of Anthropic's primary auditor becomes a contested political claim, the audit-based pillar of the Amodei proposal loses force even before any voluntary pacing agreement is reached.

What to watch next

Amodei's proposal called for international coordination and third-party evaluators with employee-level access. Whether Anthropic and OpenAI move ahead on a bilateral basis without antitrust cover is the first signal to watch. Congressional Republicans' response to any formal voluntary standards body proposal will indicate whether the political headwinds harden into active legislative opposition. Large enterprise customers and cloud hyperscalers that have not yet taken a public position on the pacing debate are a second signal: if a hyperscaler publicly backs or opposes the coalition, that shifts the commercial calculus for the labs involved.

Sources