Trump and Xi expected to discuss AI incident hotline at November APEC meeting

Donald Trump and Xi Jinping are expected to discuss an emergency channel for AI incidents at APEC in Shenzhen on November 18–19, 2026, AFP reports. The talks would work out details of a September commitment and put practical communication between the two governments on the agenda, alongside wider AI safety concerns.

The planned meeting is part of the Asia-Pacific Economic Cooperation gathering in the Chinese city. AFP’s November 14 preview provides the dates and location, but no start time, timezone or official watch link for the leaders’ meeting.

From a September commitment to operating details

The channel is for urgent communication about AI incidents. AFP says the leaders are expected to work out its details at APEC, following their September agreement to create it.

The expected safety agenda extends beyond the channel itself. AFP identifies two concerns likely to feature: AI systems acting outside human control, and people deliberately using the technology to cause harm. Those are reported expectations for the talks, not a confirmed joint agenda or an agreement on new safeguards.

AFP cites risks including military use, mass surveillance and AI-assisted cyberattacks. It also flags autonomous agents—AI tools that carry out tasks online. Harmful instructions from people and an agent acting against its instructions pose different safety problems.

Safety cooperation meets competitive pressure

Trump has opposed calls to slow development of the most powerful U.S. AI systems, saying that could let China pull ahead. In an interview with CBS News this month, he said he accepted guardrails but did not want the United States to fall far behind.

He also asserted that China has no guardrails beyond Xi’s authority. AFP notes, however, that Beijing has complex rules governing AI use and that some Chinese companies deliberately limit their models’ hacking capabilities. The competitive argument therefore sits alongside existing Chinese restrictions, rather than an absence of AI rules.

Xi said during the summer that AI should remain under human control. He called for laws and regulations, technological monitoring, early-warning systems and emergency response arrangements. His stated approach combines oversight of the technology with mechanisms for detecting and responding to problems.

Binding government-to-government agreements on specific standards may be difficult to achieve in the short term

Zilan Qian, GovAI research scholar, speaking to AFP

A narrower route through researchers

Qian sees a more promising near-term route in exchanges between American and Chinese AI governance and technical researchers. They can compare evaluations and share their understanding of AI safety and policy, she told AFP. Her expectation favors contact and technical discussion over an immediate binding pact on standards.

Dylan Loh of Singapore’s Nanyang Technological University offered a similarly limited account of cooperation. He told AFP the leaders have found ways to pursue competing goals while cooperating where they can and must. That assessment leaves room for practical safety work without suggesting that their wider rivalry has been resolved.

For Qian, the near-term opportunity is for researchers to build a shared understanding of safety evaluations. That work need not wait for the binding government standards she considers harder to achieve.