Senior Chinese and U.S. officials held an approximately eight-hour meeting in New York on Sunday. After the talks, U.S. Treasury Secretary Scott Bessent said both sides discussed setting up a communication channel called the “U.S.-China AI Dialogue,” and that the U.S. side proposed establishing an AI incident notification mechanism. The move is intended to pave the way for this week’s White House summit between Trump and Xi Jinping. （Backgrounder: Trump fires shots about slowing down AI development—points to and warns the CEO of Anthropic, accusing it of trying to disguise itself as a “perfect little angel.”） （Additional background: King Charles III convenes an AI summit! Amodei warns that within six months, AI agents will take over the human internet.）   Senior officials from the two countries held a roughly eight-hour marathon meeting in New York on Sunday. In a rare step, they brought “AI safety” to the negotiating table. Afterward, Bessent confirmed to reporters that both sides had already begun discussing establishing a communication channel called the “U.S.-China AI Dialogue,” and that the U.S. side officially proposed putting in place an AI incident notification mechanism between the two countries. Under the plan, when either side’s AI situation could rise to a national-security level, the other side would be informed immediately. The meeting is seen as a key warm-up ahead of this week’s White House summit between Trump and Xi Jinping. Led by China’s Vice Premier He Lifeng, the Chinese delegation included U.S. Trade Representative Jamieson Greer in addition to Bessent. The talks took place at the JPMorgan Chase headquarters in New York. Bessent described the day’s outcome as a “very successful engagement,” stressing that both sides agreed to meet again. 「U.S.-China AI Dialogue」: From tabletop discussion to a formal mechanism According to France 24, the “U.S.-China AI Dialogue” Bessent referred to would specifically involve creating a government-level reporting channel to handle potential AI incidents that could “escalate to a national-security level.” If it comes to fruition, it would be a rare attempt to institutionalize AI governance between the two major powers. Over the past two years, the interactions between the two sides in AI have mostly stayed at export controls, chip bans, and a race in R&D; now they are beginning to move toward the technical arrangements of “how to notify each other if something goes wrong.” It is worth noting that a newly appointed trade representative, Li Chenggang, appeared on the list of Chinese officials attending the talks, suggesting Beijing has deployed both diplomatic and trade tracks in this round. However, after the meeting, the Chinese delegation did not speak to the media, leaving the message release to the U.S. side alone. Trade proceeds in parallel: first thaw non-sensitive goods Beyond AI, trade remains the main focus. Greer said both teams are “continuously working” on the possibility of reducing certain tariffs on each other’s goods. The goal is to identify a “non-sensitive” list—China could include consumer products, while the U.S. could include energy, agricultural products, and medical devices. Continued supply of rare-earth magnets and critical minerals is also a key concern for the U.S. Officials familiar with the discussions said the talks at the working level may continue into Monday. Previously, during the negotiations, China’s commitment to purchasing U.S. agricultural products has also been part of what is under scrutiny. Meanwhile, the war between the U.S. and Iran has entered its seventh month. With China as Iran’s largest oil customer and diplomatic backer, Washington’s trade calculus toward Beijing carries an additional strategic layer. Analysts: Don’t expect too much from the Trump–Xi summit But regarding the Trump–Xi summit this week, the diplomatic community generally does not expect major breakthroughs in the short term. Kurt Campbell, a former U.S. deputy secretary of state, described that in the short run, “the only thing you can expect is a ceasefire,” and that this ceasefire “could extend to the end of this year.” He also expects Trump to announce that he will meet Xi Jinping two more times—at the APEC summit in November in Shenzhen, and at the G20 meeting in December at Trump’s Doral resort in Miami. For readers in Taiwan, the signal in this news is not the technical details of an AI notification mechanism itself, but rather the U.S. formally incorporating AI safety into the bilateral negotiation framework with China. That means AI governance in the future may be discussed alongside chip controls and supply-chain reviews. When Washington and Beijing begin to draw the boundaries for how they will notify each other about “AI incidents,” Taiwan—an intersection point for the two major semiconductor supply chains—will inevitably be pulled into the surrounding context of such discussions more often. Related report 「Yes, Mr. President」: Huang Renxun calls Trump before his 1,000-person speech agrees that the AI-slowing argument is a scam CA plans to ban AI chatbots as mental-health therapists! SB 903 passes unanimously, three readings imminent Unprecedented! Trump signs the new U.S. bills entering circulation in June, ending the Treasury Department’s 165-year tradition Bitcoin mining firm Hut 8 founder warns: AI could trigger a financial system crisis—moving funds back to BTC Trump–Xi summit kicks off today: Xi Jinping reaches out first; Trump leads with a 14-second handshake—U.S. arms sales to Taiwan, Iran, and tariffs all on the table \"U.S.-China officials hold a private 8-hour meeting in New York: propose an AI incident notification hotline to pave the way for the Trump–Xi summit\" This article was first published on BlockTempo (BlockTempo—an especially influential blockchain news media).