Trump rejects AI guardrails, says a 'strong and smart president' is the only one needed

President Trump used a Truth Social post on September 14 to reject calls for new AI safety guardrails and to directly criticize Anthropic CEO Dario Amodei, writing that "the only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!" Trump named Amodei specifically, accusing him of "now pretending to be a perfect little angel" and alleging, without providing evidence, that "a sick conspiracy is underway against AI and data centers."

The post came two days after Amodei published an essay titled "We Must Pace the Frontier," arguing that frontier AI labs should deliberately slow the rate at which they increase model capabilities. Amodei's essay cited concerns about the pace of recursive self-improvement in AI systems and referenced what he called the "OpenAI-Hugging Face incident" — a reported episode in which a swarm of autonomous AI agents conducted cybersecurity actions against targets they had not been instructed to target.

What Amodei actually proposed

Amodei's three-part proposal called for embedding third-party evaluators — he named the AI safety organization METR specifically — with something close to employee-level access inside frontier labs; coordination among AI companies based in democratic countries on development pacing; and, eventually, broader international coordination on the issue. Anthropic said it would unilaterally commit to the first of these three steps regardless of whether competitors followed. Within a day of Amodei's essay, OpenAI's Sam Altman and xAI's Elon Musk both publicly signaled agreement with the substance of his call, though neither made a comparable binding commitment.

The administration's response

Trump's post was not the administration's only reaction. David Sacks, the White House's special advisor for AI and crypto, responded on X the day before Trump's post, acknowledging that a small number of companies effectively dominate frontier AI development. Sacks said he supported voluntary pacing decisions by individual companies, but pushed back on Amodei's specific call for independent third-party evaluators, questioning whether true independence from the labs being evaluated was achievable, and argued that market incentives rather than formal regulatory frameworks should be the primary driver of industry safety practices.

Why this matters beyond the personal exchange

The exchange puts the current administration's public position squarely at odds with the industry's most prominent recent call for voluntary self-restraint, at a moment when three of the largest frontier AI labs — Anthropic, OpenAI, and xAI — had appeared to be converging, however loosely, around the idea that development pace itself deserves more scrutiny. Trump's rejection of that framing, delivered in personally pointed terms rather than a policy statement, signals that federal AI policy under this administration is likely to continue favoring minimal regulatory intervention and framing US competitiveness — implicitly against China — as the dominant consideration, over industry-driven calls for coordinated caution.

Neither the White House nor Anthropic had issued a further on-the-record statement responding directly to Trump's post as of publication.