- President Trump dismissed AI extinction fears the week of September 11-13, 2026, telling reporters in Dallas ‘No, I don’t have any’ concerns about AI causing human extinction, a direct rebuttal to public warnings from Anthropic and OpenAI researchers that same week.
- Trump estimated the US holds roughly a one-year lead over China in AI development, framing that gap as the primary national security concern and the main argument against heavy regulation.
- Six months after Trump’s December 2025 executive order aimed at curbing state AI laws, California, Colorado, Connecticut and Illinois are pressing ahead with legislation targeting algorithmic bias, employment automation and child protection.
The same week Anthropic’s alignment science lead publicly estimated more than a 10% chance that advanced AI causes human extinction within a decade, President Trump told reporters in Dallas, ‘No, I don’t have any’ concerns about it. He argued that winning the AI race against China matters far more than slowing development down to address those fears.
Trump Dismisses Extinction Warnings
Trump’s comments came from two locations: Dallas on September 11 and Ireland on September 13. In Dallas, asked directly whether he had concerns about AI leading to human extinction, he replied, ‘No, I don’t have any.’ ‘I have concerns that if we don’t win AI, we’re going to be put in a very bad position,’ he continued. ‘We are leading China right now by a pretty good period. I would say a year, which is… considered a lot.’ The remarks were a pointed counter to calls, from executives at companies including Anthropic and OpenAI for regulatory intervention or a slower pace of development.
The warnings Trump was responding to were specific. Evan Hubinger, Anthropic’s alignment science lead, posted on X estimating a greater than 10% probability that advanced AI causes human extinction within a decade. Geoffrey Hinton, a widely recognised figure in AI research, reportedly described that 10% estimate as “not unreasonable.” Jacob Coxon, a former Anthropic researcher who also worked at OpenAI, resigned from Anthropic citing concerns that companies were “gambling with our lives” by pushing self-improving AI models too quickly, and called on the US to coordinate with other countries to avoid a dangerous development race. OpenAI CEO Sam Altman and Tesla CEO Elon Musk have also publicly endorsed calls for a slower pace of development, though the two have diverged sharply on other AI questions.
The concerns these figures raise generally centre on “misalignment”, the risk that AI systems optimise for objectives that diverge from human values, and that sufficiently capable systems could be difficult or impossible to correct once that divergence occurs. Trump’s position does not engage with that framing; his concern is competitive, not technical.
The Geopolitical AI Race
Trump’s stated rationale for resisting heavy regulation is straightforward: losing the AI competition to China would, in his view, leave the US in a strategically dangerous position. “Whoever wins AI, wins,” he said. He estimated the US currently holds a lead of “roughly on the order of one year” over rivals, a gap he described as “a considerable span of time.”
That framing treats AI risk and AI competition as a trade-off, where slowing development to address safety concerns concedes ground. While Trump acknowledged that “some regulation” and “guardrails” were possible, he offered no specifics. The administration’s December 2025 executive order, discussed below, reflects the same logic: speed and national competitiveness first, with regulation confined to what does not impede either. For more on how the US-China technology rivalry shapes AI policy, see our coverage of China’s AI infrastructure build-out and its effects on the global hardware market.
Federal and State Regulatory Divide
In December 2025, Trump signed an executive order titled “Ensuring a National Policy Framework for Artificial Intelligence,” which directed federal agencies to review state AI laws deemed “onerous” and tied federal funding compliance to the order’s principles. The administration’s argument was that a patchwork of 50 different state regimes would impede both innovation and national security.
Six months on, the order has not visibly stopped state legislators. California, Colorado, Connecticut and Illinois are all advancing AI legislation, with bills targeting algorithmic bias in hiring, automated management practices and the use of manipulative chatbots aimed at children. The White House has not publicly indicated it has taken legal action against any state AI law or withheld federal funds under the December 2025 order. Whether the order has any practical deterrent effect on state legislatures, or whether the administration intends to enforce it, remains unclear.
