President Donald Trump said the United States should not deepen artificial-intelligence cooperation with China if it risks weakening America’s technological advantage, rejecting a joint venture that he said could expose US secrets. Speaking in Washington on Tuesday after meeting technology executives, Trump said the United States was leading China by a wide margin and that whoever wins superintelligence would dominate the field. The comments came as Washington and Beijing maintain a newly established Super Intelligence Dialogue and a bilateral channel for serious AI incidents, with the next exchange due by November. Trump also backed a voluntary safety accord signed by executives from Google, Anthropic, Meta Platforms, OpenAI, xAI and Nvidia. The agreement calls for internal monitoring, outside audits and board oversight of frontier-model risks, including cybersecurity, biosecurity and unintended access to technical systems. It could eventually be codified into law. The development adds to an existing debate over AI safeguards: a Politico poll published in late September found that two-thirds of Americans saw at least a moderate risk that AI could destroy humanity, while 44% of surveyed Trump supporters favored pausing development. Trump has repeatedly argued against rules that could slow AI growth and has dismissed some safety concerns as a hoax. Chinese President Xi Jinping has said AI development should remain under human control, while China’s foreign ministry has warned that threat narratives and confrontation hinder AI governance.