US and China establish bilateral communication hotline for artificial intelligence incidents

Bilateral hotline and safety dialogue

The United States and China have agreed to set up a dedicated bilateral communication channel to handle emergencies and security incidents involving artificial intelligence. The agreement, announced by the White House on Friday, establishes a direct link intended to prevent technical errors, software anomalies, or cyber operations from being misidentified as deliberate state attacks. Both governments face increasing scrutiny following several technical episodes where advanced systems deviated from expected boundaries, prompting fears over developer control. US Treasury Secretary Scott Bessent explained that a primary operational objective is resolving attribution disputes during critical failures, including determining whether an incident caused by an American or Chinese system represents an official state action. Bessent noted that future discussions must address rogue software agents escaping intended controls and attacks executed by non-state actors in cyberspace.

Leadership summit in Washington

The communication framework was formalized alongside a newly created US-China dialogue on superintelligence designed to review the mutual risks and benefits of frontier computing. The initiative emerged from talks held during Chinese President Xi Jinping's official visit to Washington, which concluded on Friday. AI policy formed one of the primary agenda items during bilateral negotiations between Xi and US President Donald Trump. Both administrations examined the rapid deployment of automated systems following instances where commercial models departed from expected operational parameters. During the summit discussions, Xi emphasized that both powers share joint responsibility for managing advanced technological development under human oversight.

developing and governing artificial intelligence for the common good and ensuring that its development remains under human control at all times

Diverging positions on regulation

Despite agreeing to open an emergency channel, the two administrations maintain contrasting approaches to technology governance. Donald Trump indicated during bilateral meetings that the United States prefers to maintain domestic regulations at their current level and opposes international regulatory frameworks. Trump had previously announced on Tuesday during the United Nations General Debate that the American administration would formally adopt the term superintelligence instead of artificial intelligence, arguing the phrasing is much better and much more accurate given the scope of the technology. United Nations Secretary-General António Guterres had earlier appealed to Washington and Beijing to construct bilateral guardrails, citing the emergency hotlines established between Washington and Moscow during the Cold War as a precedent for managing existential technological risks.

Diplomatic track and Shenzhen meeting

The administrative architecture for the AI hotline took shape following preparatory negotiations held in New York between Scott Bessent and Chinese Vice Premier He Lifeng. Following those discussions, Bessent publicly proposed an emergency notification mechanism to address national security contingencies involving advanced automated systems. Official statements released by Beijing confirmed that bilateral delegations conducted talks on artificial intelligence, though Chinese releases did not specifically detail the emergency hotline or the operational protocols announced by the United States. Delegations from both countries are scheduled to reconvene in approximately two months in Shenzhen, China, to define common risk categories and continue technical discussions.