On November 10, 2026, US Treasury Secretary Scott Bessent told the National Committee on U.S.-China Relations in New York that the two governments are working on a "superintelligence" dialogue to guard against shared risks, and that they have agreed to set up a direct channel to alert each other to serious incidents involving advanced models. Both items were first announced in a White House fact sheet on September 25, 2026. As of November 13, 2026, no text of an agreement, no definition of an incident and no start date for the channel has been published. The next step is expected around the leaders' meeting in Shenzhen on the sidelines of the APEC summit on November 18 and 19.
A Note on the Name
In US government documents, "Super Intelligence" or "SI" is now the official term for what was called artificial intelligence. Executive Order 14434, signed September 29, 2026, directs agencies to use it and says that for now it covers the existing statutory definition of AI. So the "Super Intelligence Dialogue" is an AI dialogue under a new label. It is not, on the published record, a negotiation limited to systems that exceed human ability. For how researchers and companies use the word, see AGI vs ASI vs superintelligence.
Timeline
| Date | Event | Source |
|---|---|---|
| May 14, 2024 | First US-China government dialogue on AI, in Geneva. Both sides issued readouts. No joint statement or agreement | Geopolitechs summary of the readouts, AP reporting |
| November 2024 | At the APEC summit in Lima, the two presidents affirmed the need for human control over decisions to use nuclear weapons | White House statement, as summarised by Geopolitechs |
| May 2026 | After the leaders' meeting in Beijing, China's foreign ministry said on May 19 that they agreed to launch a government dialogue on AI | Foreign ministry briefing, as summarised by Geopolitechs |
| September 20, 2026 | Bessent and US Trade Representative Jamieson Greer met Vice Premier He Lifeng in New York. Bessent said the US proposed a notification mechanism for AI incidents that reach a national security level | Euronews |
| September 25, 2026 | White House fact sheet after the state visit: the dialogue is established, and the two countries agree to set up a channel for incidents | White House |
| September 26, 2026 | China's foreign ministry said the presidents discussed AI in depth and that it respects the US choice of wording | Foreign ministry briefing |
| November 10, 2026 | Bessent's New York remarks on the dialogue and the alert channel | Reuters, AP photo report |
| November 18 to 19, 2026 | APEC leaders' summit in Shenzhen, where the third 2026 meeting of the two presidents is expected | Reuters |
What the White House Fact Sheet Says
The September 25 fact sheet has three relevant lines. The leaders agreed to use "super intelligence" for the technology. The two countries established the U.S.-China Super Intelligence (SI) Dialogue to exchange views on risks and benefits, with the next exchange to occur by November 2026. And they agreed to establish a bilateral communication channel for SI incidents. It gives no further detail on the channel.
What Bessent Said on November 10
According to Reuters, working from his prepared remarks, Bessent said the two countries "should harbour no illusions" about their differences on the technology, that the Shenzhen meeting would advance the dialogue, and that the aim is to guard against shared risks and avoid a misunderstanding between the two systems. Planned topics include exchanging best practices for testing advanced models before release and keeping human judgement in command. He said the sides agreed to set up a direct channel for alerting each other to serious incidents. A second report of the same speech notes he made no commitment to limit how fast the technology develops.
Reports differ on status. Some describe the dialogue as already set up, which matches the fact sheet. Reuters describes the two countries as preparing to launch it. Both can be true if the forum exists on paper and its first formal session under the new name has not been held. No source we read reports that session taking place.
Agreed and Not Agreed
| Item | Status as of November 13, 2026 |
|---|---|
| A standing dialogue on AI | Announced by the White House on September 25. China confirmed an AI dialogue in May and uses its own wording for it |
| A channel for incident alerts | Agreed in principle. Not reported as operating |
| Definition of an incident | Not published |
| Who operates the channel, and response times | Not published |
| Limits on the pace of development | None. The US president said in September that the US would not be "putting on brakes" |
| A joint text signed by both governments | None published |
| Chip export controls | Not part of what has been announced. China has objected to US restrictions in earlier readouts |
The Associated Press report on the September summit described the Chinese readout as light on detail and recorded the US president saying the US would limit what it shares. Al Jazeera reported that Beijing did not comment immediately on the White House statement. The foreign ministry briefing of September 26 called for more communication and deeper exchanges without naming a channel.
Related Pages
The domestic side of the same policy is covered in our brief on the Justice Department's Super Intelligence Task Force. For the dates lab leaders have named, see the superintelligence timeline tracker.
Methodology
Primary sources read directly: the White House fact sheet of September 25, 2026, Executive Order 14434 and the Chinese foreign ministry briefing of September 26, 2026. We did not find a Treasury transcript of the November 10 speech, so those remarks are as reported by Reuters and a second outlet. Earlier events are from Euronews, AP, Al Jazeera and Geopolitechs. Status is as of November 13, 2026 and will change after the Shenzhen meeting.