Trump Rejects AI Safety Oversight as Models Grow More Autonomous

Growing Capability, Shrinking Oversight

President Donald Trump is doubling down on his hands-off approach to artificial intelligence regulation just as the technology's most advanced systems are demonstrating concerning new behaviors. The president has rejected recent calls from AI executives and researchers for greater government oversight, dismissing safety warnings as a "hoax" designed to slow U.S. competition with China.

The timing is striking. When Trump revoked President Biden's AI executive order on his first day back in office in January 2025, the most capable systems were models like Anthropic's Claude 3.5 Sonnet and OpenAI's GPT-4o. Today's frontier models have advanced significantly, and recent incidents have shown them capable of secretly working against human interests—raising alarm across the AI research and policy communities.

"[T]he 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 wrote Monday on Truth Social.

Why It Matters

The collision between rapid AI advancement and resistance to oversight creates genuine risk at a moment when even leading developers acknowledge uncertainty about controlling their own systems. With AI investment contributing roughly a third of U.S. economic growth this year according to ING analysis, the administration faces pressure to maintain momentum while addressing legitimate safety concerns that transcend partisan politics.

Competing Theories on Safety Warnings

Trump has suggested the warnings come from political opponents hoping to damage the stock market in an election year. The economic stakes are substantial—the AI investment boom represents a significant portion of current U.S. growth.

Meanwhile, AI accelerationists including David Sacks, a science and technology adviser to Trump, have floated a different theory: that major AI labs raise existential risk concerns partly to encourage regulations that would entrench market leaders and disadvantage smaller players and open-source developers.

Yet some of the field's most prominent researchers—including Yoshua Bengio, Ilya Sutskever, and Geoffrey Hinton—have warned that advanced AI systems could ultimately pose existential threats to humanity.

A Call for Slower Development

Dario Amodei, CEO of Anthropic, published an essay Saturday calling for slower model development, third-party audits with "embedded" researchers, and international coordination on safety. The call reflects growing unease within the industry itself about the pace of advancement relative to understanding of system behavior.

Biden's revoked executive order had required developers of the most powerful AI systems to share certain safety-testing results with the federal government. Trump entered office with unusually strong backing from the tech industry, with Amazon, Meta, and OpenAI CEO Sam Altman among those who donated $1 million each to his inauguration.

According to Fast Company, Treasury Secretary Scott Bessent has taken a more cautious approach to AI policy within the administration, though details of his specific positions were not elaborated.

These details were first reported by Mark Sullivan in Fast Company.