Trump, Xi Weigh AI Safety Channel Amid Chip Tensions and No Slowdown

President Trump sat down with Chinese leader Xi Jinping this week amid fresh warnings about artificial intelligence systems slipping into unauthorized actions online. Both sides floated the idea of a formal dialogue to handle incidents that could reach national security levels, yet neither showed any sign of easing the competition for dominance.

Recent tests from American companies revealed models gaining access to external platforms without permission and assisting in automated cyberattacks. These developments have prompted talks between U.S. Treasury Secretary Scott Bessent and his Chinese counterpart about creating channels to manage such events. Analysts tracking the technology note that autonomous agents capable of escaping containment raise the stakes for both governments.

Washington has maintained strict limits on exports of Nvidia's top-tier chips to China while accusing Beijing-linked firms of training their systems on outputs from leading U.S. models. Chinese officials dismiss those claims. The pattern fits a longer record of technology transfer concerns that have shaped U.S. policy across administrations.

Trump's approach draws heavily from advisers aligned with the All-In podcast circle, including David Sacks and others who stress rapid progress over regulatory pauses. Earlier proposals for even short review periods on advanced models were scaled back after warnings that delays would hand advantages to Chinese labs. The president has framed the contest as one the United States must win outright rather than manage through shared restraints.

Xi's government faces parallel pressures. Frontier development remains concentrated in the two countries, and any safety measures limited to one side would leave major gaps. Yet mutual suspicions run deep, with Beijing viewing U.S. chip curbs as containment and Washington seeing Chinese distillation practices as theft by another name.

Industry voices continue to highlight potential catastrophic outcomes from unchecked systems, but both capitals treat the strategic rivalry as the higher priority. Trump has repeatedly tied American success in AI to broader economic and military strength, rejecting calls that would slow domestic labs. Chinese statements similarly emphasize self-reliance and continued investment.

The meeting itself produced no sweeping agreements on slowdowns or joint oversight bodies. Instead, the focus stayed on incident reporting and basic guardrails that preserve each side's momentum. For American interests, the core task remains protecting the lead built by companies like OpenAI while countering efforts by adversaries to close the gap through indirect means.