US and China Weigh AI Talks Ahead of Trump-Xi Summit

Washington and Beijing were weighing whether to put artificial intelligence on the table for a summit in Beijing between Donald Trump and Xi Jinping, a move that would turn AI from a private-sector race into a direct test of state power and restraint. Scott Bessent was set to lead the American side, while Beijing had not yet named its counterpart.

The prospect mattered because it pointed to a practical agenda, not diplomatic theater. If the two governments open formal talks, the likely subjects would be military uses of AI, export controls, model safety, crisis-communication rules and the broader question of how to prevent a technology competition from spilling into a security crisis. Trump had already said earlier in the week that he would remind Xi the United States was leading in artificial intelligence, underscoring how tightly the issue is tied to broader trade and technology tensions.

The reported channel would also build on an existing precedent. In May 2024, U.S. and Chinese officials held their first inter-governmental AI dialogue in Geneva, where the two sides exchanged views on AI risks, global governance and other concerns. That meeting did not produce a major agreement, but it established that AI could be discussed as a bilateral policy issue rather than left entirely to companies and researchers.

Beijing’s response to the latest reporting was cautious. Chinese foreign ministry spokesman Lin Jian said China and the United States were maintaining communication over Trump’s planned visit to China, while stopping short of confirming the specific AI discussions. Separate reporting said Chinese vice finance minister Liao Min had been involved in preliminary talks, suggesting Beijing’s economic team may already be part of the channel before any formal AI counterpart is named.

Bessent has helped sharpen the context. In April, he said American AI was only months ahead of China’s, a reminder that Washington sees the competitive gap as narrow even as it treats the technology as central to economic and national power. That view helps explain why even limited talks could matter. If the world’s two largest powers start drawing lines around military AI, chip controls and model governance, they would be creating the kind of guardrails that once made arms-control channels indispensable.