US and China negotiate AI incident reporting mechanism ahead of Trump-Xi summit

Bilateral talks in New York

On Sunday, senior officials from the United States and China met at the JPMorgan Chase headquarters in New York to negotiate frameworks on artificial intelligence, trade imbalances, and critical minerals. The delegation included US Treasury Secretary Scott Bessent, US Trade Representative Jamieson Greer, and Chinese Vice Premier He Lifeng, who also held a private bilateral discussion with Bessent. The meeting served as direct preparation for the upcoming summit between US President Donald Trump and Chinese President Xi Jinping, scheduled for Thursday in Washington. Xi is slated to visit the United States from 23 to 25 September, marking his first trip to the country since 2023 and his first reception at the White House since 2015. Chinese state news agency Xinhua described the discussions as frank, detailed, and constructive.

Proposed artificial intelligence mechanism

During the New York session, US officials presented a proposal for an early warning and incident notification mechanism to manage AI safety risks that could affect national security. Treasury Secretary Bessent explained that establishing structured communication between the two leading AI powers is intended to replace regulatory opacity with operational transparency.

We want a common vision regarding shared goals and shared threats.

Bessent advocated for guardrails to prevent advanced AI models from falling into the control of hostile non-state actors or criminal networks. On Sunday, Ro Khanna, the ranking Democrat on the House Select Committee on China, called for a formal international treaty that would ban self-improving AI systems and establish explicit prohibitions against deploying artificial intelligence in biological and nuclear weapons development. George Chen, a digital policy expert at The Asia Group, noted that future sessions between the two governments will likely expand to cover military AI deployment, critical infrastructure protection, and cyberattack prevention.

Technology competition and domestic policies

The AI discussions take place against competing commercial and technological pressures in both capitals. In August, Chinese firms including Alibaba, DeepSeek, Moonshot, and Zhipu AI released large language models capable of matching leading American software, despite US semiconductor export restrictions in place since 2022. While US executives such as Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have urged a slower pace of AI development to address safety concerns, President Donald Trump rejected calls for federal regulation.

The only reason why the AI/data center explosion is happening right now is because the United States is way ahead of every other country right now. Do not slaughter the golden goose!

In Truth Social posts, Trump compared warnings about artificial intelligence to climate concerns and reiterated that whoever leads in AI dominates global competition. In contrast, Chinese officials expressed concern that Washington is using safety talks to slow China's technological catch-up, pointing to Xi Jinping's June address at the World AI Conference where he stated that AI development must not be steered by a single country alone.

Trade balance and tariff deadlines

Beyond technology governance, the New York delegations reviewed contentious bilateral trade issues ahead of Thursday's summit. Negotiators discussed the establishment of a proposed Board of Trade as well as compliance with existing trade agreements. Trade Representative Greer stated that the US administration is seeking a more balanced trade relationship, focusing on the persistent trade deficit with Beijing. A central topic remains the bilateral tariff pause, which is scheduled to expire on 10 November. A senior US official stated that Beijing has not yet adequately fulfilled its prior commitments to facilitate the export of critical minerals, setting trade and raw materials as primary agenda items for the leaders' meeting in Washington.