Trump dismisses AI safety warnings as tech leaders call for a slowdown

At a golf resort on the Irish coast on September 13, 2026, President Donald Trump drew a clear strategic line: the United States will not slow its AI development, because ceding ground to China is a greater danger than any risk posed by the technology itself. The question is whether that line holds without a framework to back it. --- THE CONTEXT --- Trump's remarks at Doonbeg followed an essay published Saturday by Anthropic CEO Dario Amodei, which called on companies and governments to reduce the speed of AI capability improvements. Amodei cited risks including loss of human control over AI systems and the technology's potential exploitation in cyberattacks and bioterrorism. --- THE FACTS --- Amodei proposed international cooperation on AI governance and the embedding of third-party evaluators inside AI laboratories, pledging that Anthropic would act unilaterally. Elon Musk and OpenAI CEO Sam Altman both endorsed Amodei's concerns on social media. Altman told Fortune Magazine that he would delay OpenAI's public offering until at least 2027, citing safety, and that officials at the company had discussed pausing certain tests. Trump told reporters in Doonbeg that warnings about AI going rogue were exaggerated and attributed them to «negative forces» promoting claims about things that would not happen. He did not define that phrase. House Speaker Mike Johnson, on CNN's State of the Union, rejected a moratorium, warning that China would overtake the United States if Congress moved to regulate AI through emergency legislation. Johnson called instead for tech companies to lead on safety. David Sacks, co-chair of the President's Council of Advisors on Science and Technology, wrote on X that he supports the voluntary decisions of Altman and Amodei but told both executives to stop suggesting they need outside permission. --- THE POSITIONS --- Trump and Johnson treat the issue as a matter of national security and competitive order: regulatory delay is a strategic gift to China. Amodei, Altman and Musk treat it as a structural risk to institutional control. Sacks occupies a middle position, endorsing voluntary caution while rejecting mandated regulation. --- WHAT REMAINS UNKNOWN --- This source does not specify what concrete safety mechanisms the administration considers adequate, nor does it establish whether a formal coordination structure between the federal government and AI laboratories currently exists. --- UNANSWERED QUESTIONS --- • What specific «guardrails» does the Trump administration have in mind, and are they binding or advisory? • Does Sacks's social-media position reflect official White House policy or personal judgment? • If voluntary restraint by companies is the government's preferred mechanism, what accountability structure enforces it? • How does the intelligence community assess the concrete advantage China would gain from a U.S. development pause? --- EPM ANALYSIS --- The administration's strategic logic is coherent: AI leadership is a national security asset, and unilateral deceleration carries real geopolitical cost. But coherent logic is not the same as a policy. Trump spoke of guardrails without specifying any. Johnson warned of China without detailing a governance alternative. The market and allied governments are left to read between lines that do not yet exist. 📌 📌 EPM Take: In EPM's view, the geopolitical argument for maintaining AI momentum is serious and should not be dismissed. China's ambitions in this space are documented and consequential. But the administration cannot have it both ways: claim leadership, then hand safety oversight entirely to the private sector without any institutional mechanism. That Altman and Musk — figures aligned with this administration — are raising the same alarm as Amodei suggests the risk is real. Dismissing them as part of unnamed «negative forces» is not a strategy; it is a liability.

---