US-China Super Intelligence Dialogue and Incident Channel Plan

A dialogue, a channel and a shared word

The White House fact sheet closing Xi Jinping’s state visit lists three AI outcomes. The two countries established a U.S.-China Super Intelligence (SI) Dialogue “to exchange views on risks and benefits related to SI,” with the next exchange due by November 2026. They agreed to establish a bilateral communication channel for SI incidents. And the two leaders agreed to use the term “super intelligence” rather than “artificial intelligence” for the technology. [1]

The Associated Press described the same result as a communication mechanism for AI-related incidents, with an AI-specific dialogue scheduled for November, alongside agreement to accelerate work on military crisis communications. The leaders also agreed to meet again at the APEC summit in Shenzhen in November. [2]

The groundwork came first. On September 20 in New York, Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng, who lead their countries’ economic and trade talks, held what China’s commerce ministry called their first dialogue on AI. Bessent said afterward that the two sides had agreed to hold further talks about a hotline to notify each other of AI incidents with national security implications. The summit turned that discussion into an agreement between the two leaders. [5][6]

The fact sheet is dated September 25, the last day of the visit, and the White House posted it late that evening Eastern time. The AP report, published the next day, attributes the outcomes to statements both governments made on Saturday, September 26, when Xi returned to Beijing. The Atlas dates the agreement to the day the White House published it. [1][2]

What each side said about AI

China’s readout of the leaders’ talks on the morning of September 24 has Xi saying the two countries “may have competition, but they may cooperate even more,” that they can continue their dialogue, exchange views on risks and benefits and “jointly prevent the misuse and abuse of AI,” and that “AI must be kept under human control.” It quotes Trump as saying AI “concerns the future of humanity” and that the two countries should maintain dialogue and strengthen cooperation. The Chinese text says “AI” throughout, not “super intelligence.” [3]

Trump’s public remarks set the limit. Speaking to reporters outside the White House, he said the United States “is not going to be putting on brakes,” that it was leading China “by a lot,” and that he would rather not integrate the two countries’ work because “when you’re leading, you don’t open it up to each other.” [2]

No limits, tests or verification

Neither government’s record commits either country to any restriction on AI development, any pre-deployment testing, or any way to verify the other’s claims. The fact sheet establishes the dialogue and agrees to establish the incident channel, but publishes no membership, procedures or definition of an incident. As of the end of September the channel had been agreed rather than shown to operate. [1][3]

Atlas interpretation: Writing for the Carnegie Endowment, Jeffrey Prescott said the leaders’ meeting yielded “no further details or apparent progress” beyond the dialogue and hotline announced beforehand, and suggested the hotline might amount to an arrangement for Bessent and China’s vice premier to pick up the phone in a crisis. That is one analyst’s reading of an arrangement neither government has described in detail. Fortune’s Jeremy Kahn raised a different limit: a notice that a rogue system is loose does not tell the receiving country what to do about it. [4][5]

Atlas interpretation: The shared term turned up again four days later, when Trump ordered federal agencies to write “Super Intelligence” instead of “AI”. China’s own readout did not adopt the word. [1][3]