The US and China: a "superintelligence" dialogue and an incident channel
During Xi Jinping's state visit to Washington on 23-25 September 2026, Donald Trump and Xi established a bilateral dialogue on the risks and benefits of advanced AI, with the next exchange in November 2026, and agreed to create a communication channel for incidents. According to a White House fact sheet of 25 September, the leaders agreed to call the technology "super intelligence" and the dialogue the US-China Super Intelligence Dialogue; the list of eight deliverables issued by China's Foreign Ministry on 26 September says nothing of such an agreement and calls it the China-U.S. AI Dialogue.
Why it matters
The two states with the largest AI programmes agreed on a standing bilateral format on the risks of frontier systems and on a channel for incidents, choosing a bilateral arrangement where two days earlier the United States had rejected a global one at the Security Council.
The dialogue is the seventh of the eight points on the Chinese list; China's Foreign Ministry puts the next exchange "in November", the White House "by November". The talks took place on the morning of 24 September: according to the Chinese readout, Xi said the two sides could continue their dialogue on AI and that AI must be kept under human control. The day after the fact sheet Trump, according to the Times of India, told reporters he did not want to "integrate" efforts with China because the United States is "leading by a lot". What the record does not claim: that the incident channel already exists; both sides say only that they agreed to establish it. No document names the day the agreement was reached, so the record spans the whole visit.