US Pitches China on AI Notification System in NYC Talks

The United States has proposed an AI notification system that would create an early-warning channel with China for artificial intelligence incidents with national security implications, a first for the two rival powers. Treasury Secretary Scott Bessent floated the idea on Sunday during hours-long talks with Chinese Vice Premier He Lifeng at JPMorgan Chase headquarters in midtown Manhattan, according to the New York Post.

The meeting took place on the eve of the United Nations General Assembly high-level week in New York, and four days before President Donald Trump hosts Chinese leader Xi Jinping at the White House for a state visit on Thursday. It marked the first in-person session between the two economic chiefs since May.

A Hotline for the AI Race

The proposed AI notification system would let Washington and Beijing alert each other to AI-related developments with national security implications, reported by the New York Post as modeled on the Washington-Moscow hotline that linked the Pentagon and the Kremlin during the Cold War. The exact design of the channel has not been disclosed, and neither side has confirmed whether Beijing accepted the concept.

During the talks, Bessent stressed American concerns that increasingly autonomous AI agents could go rogue or be turned toward developing biological or other weapons of mass destruction, according to the report. After the meeting, Bessent told the Financial Times that simply opening a sustained dialogue was the main achievement, noting that the two sides established an AI dialogue and expect to meet again within two months.

The proposal surfaced only days after several leading AI companies publicly called for a global slowdown in AI development, arguing that the technology was advancing faster than safety systems could manage. Trump has rejected that demand, warning that an enforced pause would hand Beijing an advantage in the race to dominate the sector. Senator Bernie Sanders has taken the opposite line, calling for Washington to cooperate with China as leading labs push toward systems that can recursively improve themselves. The contrast highlights a split in American politics between those who see speed as security and those who see cooperation as the safer bet. Notably, the push for an alert channel follows an AI self-regulation debate in Europe, where Spanish Prime Minister Pedro Sanchez recently argued that self-regulation had failed and pressed for binding oversight.

Trade Truce Left Unresolved

For all the movement on AI, the two sides punted on the question of extending their fragile trade truce, which is set to expire on November 10. Washington suggested a half-year extension, while Beijing pushed to lock the arrangement in for the remainder of Trump's term, according to the Financial Times. Bessent told the paper that both sides want to preserve the stability of the past year, but no long-term deal emerged from the Sunday session.

The truce followed one of the sharpest trade confrontations in modern history. Trump opened his second term by imposing triple-digit tariffs on Chinese goods, prompting Beijing to retaliate in kind. Both capitals later stepped back from the brink and formalized a de-escalation agreement last November, but the underlying disagreements over market access, technology controls and industrial subsidies were not fully resolved. The ticking expiration date gives Thursday's summit added stakes.

All Eyes on Thursday's Summit

The diplomatic choreography in New York this week shows how much is riding on the Trump-Xi meeting. On Monday, Secretary of State Marco Rubio huddled with the foreign ministers of South Korea and Japan in a New York hotel, reported by Reuters, as both allies voiced quiet anxiety that Trump could offer concessions to Xi on Taiwan or trade in exchange for an economic headline. Japanese Prime Minister Sanae Takaichi and South Korean President Lee Jae Myung are both eager for face time with Trump during the General Assembly, where all three leaders speak on Tuesday.

Seoul is also negotiating the details of a large investment package in the United States tied to last year's bilateral trade deal, a matter that drew parliamentary scrutiny at home this week after an opposition lawmaker warned that energy project negotiations could leave South Korea carrying outsized financial risk. Takaichi, who angered Beijing last year by suggesting Tokyo could be drawn into a conflict over Taiwan, has likewise been pressing for Trump's ear before the summit.

Trump's Thursday session with Xi will be the Chinese leader's first visit to American soil since 2023. Whether the AI alert proposal advances from a conversation into a mechanism may not be known until after that meeting, but the fact that the United States raised it at the Treasury level signals that AI safety has moved from a research-lab talking point to a fixture of great-power diplomacy. For now, the proposal remains just that: a proposal, floated in a Manhattan conference room four days before two of the world's most powerful leaders sit down in Washington.