US proposes AI crisis channel with China ahead of expected Trump–Xi summit

The United States has proposed a dialogue and notification mechanism with China for serious artificial intelligence incidents that could threaten national security, as Washington and Beijing compete for technological advantage.

US Treasury Secretary Scott Bessent discussed the proposal with Chinese Vice Premier He Lifeng during talks in New York, according to Al Jazeera. Bessent said the framework would address “common goals and common threats” and move the two countries towards greater transparency.

The proposal’s operating details remain unclear. Neither the incidents that would trigger a notification nor the mechanism’s governance or permanence have been disclosed. China’s official Xinhua news agency confirmed that the officials discussed artificial intelligence but did not say whether Beijing had accepted the proposal.

The initiative would add an AI-focused channel to existing US–China crisis communications, including a military hotline established in 2008. That arrangement was intended to allow senior defence officials to communicate during dangerous incidents, although US officials have previously complained that calls were not always answered.

The proposed mechanism comes as both countries pursue AI leadership while facing calls for safeguards around rapidly advancing capabilities. The United States retains advantages in private AI investment and advanced semiconductors, while China has strengths in rare-earth supplies, electricity generation and the development of models despite restrictions on access to leading chips.

The issue may be raised at an expected Trump–Xi summit in Washington this week, although the meeting’s arrangements and agenda remain subject to confirmation. Trade, tariffs, Iran, rare earths and Taiwan are also expected to feature in the broader discussions.