The artificial intelligence race between the United States and China is taking a new turn as President Donald Trump and Chinese President Xi Jinping prepare to discuss AI safety and national security risks at their upcoming summit. The talks come as both countries push aggressively to strengthen their positions in advanced AI, even as concerns grow over increasingly autonomous systems capable of taking actions with limited human oversight. U.S. Treasury Secretary Scott Bessent has already discussed the possibility of establishing a U.S.-China AI dialogue with Chinese Vice Premier He Lifeng, including a notification mechanism for AI incidents that could reach the national-security level.

The push for dialogue comes against the backdrop of growing concerns about what powerful AI systems could do if they operate beyond expected controls. Recent incidents involving AI systems and cybersecurity have intensified calls for greater safeguards, with experts warning that autonomous AI agents could make cyberattacks faster and more difficult to contain. For Washington and Beijing, the challenge is finding areas where they can communicate on AI risks without slowing their broader race to develop and deploy increasingly advanced technology.

Despite the potential for cooperation, major disagreements remain. The United States continues to restrict China’s access to some advanced semiconductor technology used to power AI systems, while American officials have also accused Chinese AI companies of using “distillation” to gain capabilities from more advanced U.S. models. China has rejected those allegations. At the same time, Chinese AI models have continued to make significant progress, while access to advanced computing power remains a major issue in the broader U.S.-China technology relationship. Reuters has also reported that the United States has approved some Chinese companies to purchase Nvidia’s H200 AI chips, although deliveries had not yet been made at the time of its report.

Analysts cited in the discussions say the most realistic areas of progress could involve agreeing on basic AI safety principles and creating an emergency communication channel for serious incidents. Such an arrangement would allow both governments to exchange information when an AI-related event poses a national-security concern, without requiring either side to abandon its technological ambitions. The U.S. proposal for an AI incident notification system has already emerged from preparatory talks ahead of the Trump-Xi meeting.

For businesses, investors and technology companies watching the global AI race, the summit could therefore produce an important signal about how Washington and Beijing intend to manage AI risks while continuing to compete. Any agreement on communication or safety standards would represent a step toward reducing the risk of misunderstandings around powerful AI systems, but disputes over semiconductor access, technology transfers and the competitive use of AI are likely to remain central to the relationship. The outcome of the discussions could shape not only U.S.-China technology ties but also the wider direction of global AI governance.

source: cnbc