Bessent Proposes U.S.-China AI Safety Incident Notification Channel in New York Talks With He Lifeng

After talks in New York on Sunday, September 20, 2026, Treasury Secretary Scott Bessent said the United States had proposed to Chinese Vice Premier He Lifeng a new U.S.-China AI dialogue including a notification mechanism for AI-related incidents that rise to a national-security level, for President Trump and President Xi Jinping to consider at their Washington summit later that week. “We think that just like with any cross-border activity, that moving from opaque to more transparency between the number one and the number two AI powers in the world is very important,” Bessent said, per NBC News. NBC reports the two sides agreed to meet again on the AI issue.

This entry records a proposal, not an agreement. Tier-1 reporting (NBC, CNN) consistently says “proposes.” One lower-tier outlet described an agreed formal channel reached after eight hours of talks, and a secondary site named a “USA-China AI Dialogues” with a Shenzhen follow-up; neither detail is in the tier-1 sources, so neither is asserted here. The capture ledger dated this September 21, the date of the NBC article; the talks were September 20. Whether anything was formalized should be read from the summit’s outcome documents.

Structurally, it routes AI safety through great-power bilateralism. Two days later Trump told the UN the U.S. “totally rejects” any “globalist scheme to control” AI (2026-09-22–trump-unga-renames-ai-super-intelligence-rejects-global-ai-governance). A Treasury-led, security-scoped channel between two governments is the governance form compatible with that posture: it covers incidents between states, not how either state uses AI against its own population.

Open: the scope of any mechanism adopted at the summit (strategic/military systems only, or domestic government deployments), and whether it is written down.