AI will be a key US-China talking point at Trump-Xi summit 

For years, the US-China technology rivalry has largely been about who gets to build, control, and deploy the most powerful artificial intelligence systems. This week, the world’s two biggest AI powers are preparing to discuss something rather different: how they might prevent the technology from creating a crisis neither side intended. 

Chinese President Xi Jinping will visit the United States from September 23 to 25 for a state visit at the invitation of President Donald Trump, with the two leaders scheduled to meet in Washington on September 24. The summit will cover the broader relationship between the two countries, but artificial intelligence has emerged as an increasingly important item on the agenda.  

Ahead of the summit, US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng held lengthy discussions in New York. Washington proposed establishing a US-China AI dialogue that could include a notification mechanism for incidents serious enough to raise national security concerns. Bessent argued that greater transparency between the world’s two leading AI powers could help both countries deal with common risks.  

Also read: AI Safety in Focus as Big Tech Debates What Comes Next 

The proposal is significant because AI competition between Washington and Beijing remains intense. The two sides disagree over advanced semiconductor access, technology transfers, intellectual property, AI regulation, and the potential military applications of increasingly autonomous systems. US restrictions on sophisticated AI chips and semiconductor manufacturing equipment have also become a major point of friction, although those export controls were reportedly not part of the latest discussions on the proposed AI mechanism.  

The immediate ambition therefore appears relatively modest. Neither side is suddenly abandoning its technological competition. Instead, the focus could be on creating channels through which Washington and Beijing can communicate if AI systems are involved in a major security incident, cyber operation, or other potentially destabilising event. 

That becomes particularly important as AI moves deeper into military and critical infrastructure systems. US and Chinese security experts have separately argued for measures including human control over consequential military decisions, safeguards around nuclear systems, and direct communication channels for AI-related emergencies.  

Still, AI will be only one piece of a much larger summit. 

Trade is likely to dominate the economic discussions. The two countries are expected to examine whether their existing tariff truce can be extended, while Washington is seeking greater Chinese purchases of American goods and Beijing wants relief from some technology restrictions. Critical minerals and rare earths are another important issue, given China’s dominant position in the global supply chain. 

Taiwan, China’s relationship with Iran, fentanyl precursor shipments, and broader military tensions are also expected to enter the conversation. The breadth of the agenda means a sweeping breakthrough would be difficult to achieve in a single summit. 

The symbolism, however, will be difficult to miss. Xi’s September 23-25 visit marks his first White House visit in a decade, while Trump is reportedly planning an unusually prominent welcome, including a possible airport reception.  

What should the world expect? Probably incremental agreements rather than a reset of the relationship. An AI communication channel, progress on trade, or an extension of the tariff truce would each represent movement. The deeper US-China competition over chips, computing power, military technology, supply chains, and AI leadership is unlikely to disappear. 

The most consequential development may therefore be less about who wins the AI race, and more about whether the two countries can agree on rules for what happens when that race becomes dangerous.