What happened
The White House's September 25 account of President Xi Jinping's Washington visit reports that the United States and China established a bilateral dialogue on AI risks and benefits, with a further exchange due by November. Separately, the governments agreed to establish a communication channel for AI incidents. Reuters reported the agreement on September 26 and corroborated it through statements from both governments. The leaders' agreed terminology is “super intelligence.” Establishing the dialogue and agreeing to create the incident channel are different stages: the available accounts do not show that the channel is staffed, tested or operating.
The potential users are the two governments' responsible agencies, which could use a working channel to notify one another of cross-border incidents and coordinate responses. The announcement does not establish incident definitions, reciprocal notification duties, access to company evidence, independent investigation or public reporting. A senators' letter dated September 23 and published September 24 sought a broader agreement on testing, monitoring, human control and verification. Record that letter as supporting advocacy, not a second operating institution. Earlier September 20–21 dialogue reporting is a precursor to this announcement, not a separate score.
Potential direction of power
Undetermined
An operating incident-notification channel could give public institutions evidence beyond company-controlled disclosures. Establishing a diplomatic dialogue alone does not demonstrate that such a channel operates.
Window of contestability
Unclear
Unclear until the incident channel operates with identifiable duties, access and responsibilities. Establishing a dialogue alone does not establish these effects.
