Beijing and Washington agreed to establish a formal China-U.S. AI Dialogue during President Xi Jinping’s September 23-25 US state visit to the United States, according to China’s Ministry of Foreign Affairs. The dialogue will provide a forum to discuss the risks and benefits of artificial intelligence. The first exchange is scheduled for November 2026, but no further details, such as participating agencies or officials, have been disclosed.

The countries will establish a bilateral communication channel for AI-related incidents, providing officials with a direct way to communicate when AI systems impact both nations.

The Chinese readout does not clarify what constitutes an “AI incident,” how the channel would function, which agencies would be involved, or whether it would address cybersecurity, autonomous systems, or military applications.

US-China AI dialogue follows growing debate over AI risks

The AI discussion between Xi Jinping and Donald Trump comes amid growing concerns in both countries about the risks posed by increasingly capable AI systems. During their September 24 meeting at the White House, Xi said the US and China had strong AI capabilities and should continue dialogue on the risks and benefits of AI while working together to prevent its misuse. Trump also said AI “concerns the future of humanity” and that the two countries should maintain dialogue and strengthen cooperation on AI.

The discussions followed Anthropic CEO Dario Amodei’s recent essay, “We Must Pace the Frontier”, in which he called for slowing the pace of AI development so that safety measures can keep up with advances in AI capabilities. Amodei proposed a framework involving independent evaluators, coordination among AI companies in democratic countries and, eventually, global cooperation. He also called for tighter controls on advanced AI chips and semiconductor equipment reaching China, stronger measures against unauthorised model distillation and better security against model-weight theft.

Amodei argued that such measures could preserve the US lead over China and create more room to develop AI safely. He wrote that “chips will be the main determinant of China’s AI strength” and said the US should use any lead it maintains to slow the pace of AI development rather than allow competition to accelerate it.

China criticised Amodei’s proposal. Foreign Ministry spokesperson Guo Jiakun called for an “open, inclusive and benevolent approach to AI” and said confrontation and malicious competition would disrupt global AI governance. The state-backed Global Times described Amodei’s proposal as a “Cold War” tactic aimed at restricting China’s AI development, while the People’s Daily called for US-China cooperation on AI risks and said AI should not be a “monopoly of great powers.”

US restrictions on advanced AI chips to China

Washington began restricting advanced AI chip exports to China in October 2022, when the Biden administration imposed broad controls on semiconductors and chipmaking equipment to prevent China’s military from accessing advanced AI hardware. These controls were expanded in 2023 to cover chips like the H800 and H200, and again in January 2025 with the “AI Diffusion Rule,” which limited global access to AI model weights and compute. The Trump administration rescinded the diffusion rule in early 2026, tightened other controls, and shifted from a presumption-of-denial licensing policy to case-by-case reviews.

Government response to DeepSeek’s R1 model release

DeepSeek’s R1 model was released in January 2025, at the end of the Biden administration. Significant U.S. government action against DeepSeek began under the Trump administration. In February and March 2025, bipartisan lawmakers introduced bills such as the “No DeepSeek on Government Devices Act,” and the Trump administration considered a federal ban on DeepSeek for government devices. Around the same time, states including Texas, Alabama, and Arkansas independently blocked the app from state devices and networks, citing concerns similar to those raised about TikTok, such as code that could transmit data to Chinese state-linked telecom infrastructure.

Also read: