During a series of high-level meetings in New York, Treasury Secretary Scott Bessent presented a plan to create a formal AI safety dialogue between the United States and China. The proposal calls for a rapid-response channel that would allow each government to inform the other when an artificial-intelligence incident threatens national security. The idea was raised in a conversation with Chinese Vice Premier He Lifeng and comes just four days before President Donald Trump is set to host President Xi Jinping in Washington.
How the AI incident hotline would work
The suggested mechanism focuses on events that are serious enough to merit cross-border attention – for example, a large-scale AI-assisted cyber intrusion, an autonomous system that behaves unexpectedly, or the misuse of a powerful model by hostile actors. While officials have not released a precise trigger list, they emphasized that moving from opacity to greater transparency between the world’s two leading AI powers is essential. In practice, the channel would convey what happened, the intelligence gathered, and the remedial steps being taken, allowing both sides to assess the situation before assumptions harden into retaliation.
Why the proposal matters amid a fierce AI rivalry
The United States has repeatedly warned that it cannot afford to let China overtake it in the AI race. Treasury officials earlier this month stressed that “if they were to pull ahead of us on AI, then nothing else matters.” President Trump has dismissed calls for a pause on frontier AI development, arguing that the winner of the competition will dominate future economic and military landscapes. By contrast, the notification system is framed as a safety net rather than a limitation, aiming to prevent accidental escalation when a disruptive AI event occurs.
Parallel progress on trade: the new Board of Trade
Alongside the AI talks, U.S. Trade Representative Jamieson Greer announced that both governments have agreed to operationalize a newly-created Board of Trade. The board is intended to manage bilateral commerce and to identify “non-sensitive” product categories that could qualify for tariff relief. Potential items include consumer goods, agricultural produce, energy commodities and medical devices. Although no specific tariff schedule has been released, the framework signals a willingness to lower barriers on a limited set of goods while keeping strategic technologies, such as advanced chips, under existing export controls.
Chinese officials, including Chief Trade Negotiator Li Chenggang did not comment on the AI notification proposal, and state media only reported “frank, in-depth and constructive exchanges on important economic and trade issues.” The absence of a detailed public agreement on the AI channel underscores that the initiative is still in an exploratory phase, with a follow-up dialogue expected within a couple of months.
Should the hotline become operational, it would join a family of emergency communication lines that major powers have used since the Cold War – hotlines for nuclear, maritime and diplomatic crises. The novelty here is the focus on algorithmic threats that can spread faster than traditional weapons. A transparent, rapid-share system could give both capitals the chance to clarify intent and prevent misinterpretations that might otherwise spark a larger confrontation.
In the broader context, the AI safety proposal and the revived Board of Trade illustrate a rare convergence of competition and cooperation. While the United States and China continue to vie for dominance in chips, data centers and research talent, they also share a mutual interest in avoiding a scenario where an uncontrolled AI incident damages critical infrastructure on either side. The coming Trump-Xi summit will likely test whether these parallel tracks can move from dialogue to concrete policy.
