Washington proposes an AI incident notification mechanism to Beijing, days before the Trump-Xi summit

Brasília, September 21, 2026. Brazil views Sunday's meeting in New York between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng as a precursor to a technological order being shaped without it. According to Estadão, the two delegations discussed for approximately eight hours at JPMorgan Chase's headquarters before Thursday's scheduled summit in Washington between Donald Trump and Xi Jinping. Bessent described the day as "muito bem-sucedida" and announced that "the United States has proposed a notification mechanism between the two countries," with the goal of sharing "a common vision of common objectives and threats" related to artificial intelligence.

For the Brazilian press, the stakes go beyond bilateral diplomacy. G1 emphasizes that Washington and Beijing consider technological advancements as strategic for their economic and military competitiveness, and that the two capitals remain divided over Chinese access to advanced semiconductors, accusations of technology transfer, and how to regulate AI. These unresolved disagreements are to be revisited during a future meeting already announced by the two delegations.

Brazil, which is preparing its own artificial intelligence law and depends on both Chinese markets for its agricultural exports and American technologies for its digital infrastructure, is closely watching how a bilateral notification mechanism could, in fact, set security standards for the entire sector without consulting third-party countries. US Trade Representative Jamieson Greer deemed it "imperative" that Washington and Beijing "be able to work together" on these issues — a cooperation that, if realized, would outline a framework of reference that emerging powers like Brazil would have to adapt to afterwards.

Thursday's summit is also expected to address the extension of the trade truce, Chinese restrictions on rare earths and magnets, and US export controls in the cloud and semiconductors — all files in which Brasília has direct interests through its own exports of critical minerals. The two delegations agreed to open a formal dialogue on AI and meet again, without setting a specific schedule or clarifying whether other countries would be involved in the discussions.

Beijing, Monday, September 21, 2026. The Chinese press is addressing the American proposal for a notification mechanism on artificial intelligence-related incidents as one element among others in a package of negotiations whose center of gravity remains commercial. On Sunday in New York, US Treasury Secretary Scott Bessent, Trade Representative Jamieson Greer, and Chinese Vice Premier He Lifeng held their "final preparatory discussions" before the summit between Xi Jinping and Donald Trump scheduled for next week in Washington, reports the South China Morning Post. The two delegations are to "initiate their long-standing dialogue on artificial intelligence security and governance" - a formulation that puts the notification mechanism in a broader framework than a simple bilateral alert on incidents.

The South China Morning Post emphasizes that the economic agenda carries at least as much weight: renewal of the Busan agreement on rare earths, and the possible establishment of a "US-China Board of Trade", a reduced tariff mechanism for non-sensitive sectors. The New York meeting reproduces, according to the newspaper, the pattern of May, when Bessent and He had met in Seoul before the Xi-Trump summit in Beijing.

The official agency CGTN insists on the institutional framework: the delegation led by He Lifeng, a member of the Political Bureau of the Central Committee of the Party, was announced by the Chinese Ministry of Commerce, "guided by the important consensus reached by the two heads of state". No Chinese statement on the precise content of the artificial intelligence notification mechanism appears in the articles consulted.

On the technological front, the South China Morning Post relays a distinct concern: potential American restrictions on cloud computing would constitute "the most severe bottleneck" for China's AI sector since the ban on advanced chips, pushing companies to rent servers abroad, in Southeast Asia or the Middle East. This risk of technological blockage, more than the notification mechanism, structures the Chinese reading of the AI file.

Finally, the province of Zhejiang, the country's second-largest exporter, has asked its industrialists to accelerate deliveries to the United States during this "commercial détente" deemed "favorable", a sign that Beijing is seeking above all to secure tangible economic gains before Thursday.

Helsinki, September 21, 2026. In the face of the announcement of a bilateral notification mechanism in the event of a serious incident related to artificial intelligence between Washington and Beijing, the Finnish press is measuring the gap with the debate that is stirring in its own country about the regulation of this technology.

US Treasury Secretary Scott Bessent indicated on Sunday, September 20, 2026, after a full day of discussions in New York with Chinese Vice Premier He Lifeng, that "the United States has proposed the establishment of a notification mechanism between our two countries" (YLE Uutiset). The goal is to exchange alerts in the event of a serious incident related to AI and to mutually warn each other of threats affecting, including, national security. Bessent described the meeting as "very successful", while US Trade Representative Jamieson Greer presented this meeting as preparation for the upcoming summit between Donald Trump and Xi Jinping in Washington.

This bilateral channel, seen from Finland, contrasts with the position displayed by Trump on American soil. The President wrote on Truth Social that the growth of the AI sector should not be slowed down "in any way", rejecting calls for strengthened regulation from the leaders of OpenAI, Anthropic, and Google, and announcing the creation of "AI forces" (tekoälyjoukot) as well as the appointment of an "AI tsar", without specifying the contours (Helsingin Sanomat, MTV Uutiset).

This contrast is fueling a already lively debate in Helsinki. A survey conducted this summer by YLE shows that more than half of Finns believe that AI will cause mass unemployment, a concern that has relaunched proposals for taxation of work done by machines, mentioned by Bill Gates, Bernie Sanders, or economist Robert Shiller. Professor Kaisa Kotakorpi, from the University of Tampere, recalls that the scientific literature on the fiscal impact of AI remains embryonic. Meanwhile, researchers from Aalto and Tampere, solicited by YLE, are listing risk scenarios: facilitated cyberattacks, targeted disinformation, and even assistance in the manufacture of biological weapons.

For the Finnish press, Bessent's announcement illustrates how AI security standards are being negotiated between two capitals, without third countries, including Finland, having a direct voice in this emerging dialogue between Washington and Beijing.

Paris, September 21, 2026. A day of negotiations in New York, at JPMorgan's headquarters, between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng: the French press sees it mainly as the prelude to the Trump-Xi summit scheduled for Thursday in Washington. "The United States has proposed setting up a notification mechanism between our two countries," Bessent said at the end of the meeting, specifying that it would be to alert each other "in case of incidents, if something has happened that affects national security."

The cited articles put this proposal in the context of technological rivalry rather than the details of its mechanics. Rivals in AI, Washington and Beijing are said to be "increasingly concerned about the risks after several episodes that have raised fears that the technology is escaping its creators." The two powers are described as engaged in "an AI race that resembles a duel": on the American side, Chinese competition serves as justification to accelerate without limits; on the Chinese side, Beijing presents itself as an alternative to American dominance over technology. Bessent summarized the stated ambition: "We think it's very important to move from a situation of opacity to more transparency between the world's first and second powers in AI." He added that he wants to "define a common vision of our objectives and the threats we face."

These discussions, described as "crucial" to prepare for Xi Jinping's state visit, were held in the presence of US Trade Representative Jamieson Greer. He Lifeng, presented as China's chief negotiator for economic and trade issues, is to continue the exchanges before Thursday's summit, which will also address the expiring trade truce in November and Chinese restrictions on rare earths. The French coverage remains focused on these two meetings, in New York and then in Washington, without detailing the precise technical content of the future notification mechanism, its implementation schedule, or how it would fit in with US export controls in the cloud and semiconductors, briefly mentioned as the backdrop to technological tensions between the two capitals.

New Delhi, September 21, 2026. The Indian press first notes that this round of discussions between Washington and Beijing, held on Sunday in New York with Treasury Secretary Scott Bessent, Chinese Vice Premier He Lifeng, and US Trade Representative Jamieson Greer, took place without India having a voice, even though its interests are directly at stake. Bessent described the meeting as "very successful", stating that they had "completed what was left pending in Beijing" during the May summit between Donald Trump and Xi Jinping. Xinhua, on the other hand, spoke of "frank, in-depth, and constructive" exchanges, guided by "mutual respect" and "peaceful coexistence".

The two delegations agreed to open a formal dialogue on artificial intelligence and the operationalization of a "trade council" outlined in Beijing in May. Washington proposed a bilateral notification mechanism for incidents related to AI affecting national security - a channel that Bessent presents as a step towards "more transparency" between "the world's two leading AI powers". The Times of India emphasizes that this framework remains strictly bilateral: neither India nor any third country is included, even though the norms being drawn up will shape global regulation of the technology.

However, New Delhi is not just a spectator: the new US sanctions law, signed by Trump in parallel, imposes 100% customs duties on countries buying Russian oil and gas - a list that explicitly includes China and India. The Hindu Business Line also notes the announcement, the previous day, of an "AI Force" and a future "AI Tsar" to maintain America's lead over Beijing.

The Free Press Journal broadens the perspective to the social consequences of the AI race on the Chinese side: according to IDC, the proportion of Chinese industrial enterprises using AI models has risen from less than 10% in 2024 to nearly half today, with job losses such as that of Beijing programmer Fei Zhaojun, who was laid off with 160 colleagues. For the Indian press, this diagnosis serves as a warning: the AI governance standard is being negotiated between two parties, but its effects far exceed the two capitals, reaching as far as New Delhi, just days before the Trump-Xi summit scheduled for September 23 in Washington.

Rome, Monday, September 21, 2026. The Italian press first retains a warning from an economist rather than an official diplomatic announcement. In La Repubblica, Gita Gopinath, an economist at Harvard and former number two at the IMF, describes a world that "is leaving things to chance" in the face of the risks of artificial intelligence: "We are relying on luck, and so far, we have been lucky. But we cannot assume that this will last forever." She cites the "OpenAI-Hugging Face incident" as a warning signal and believes that policymakers are being outpaced by the speed of AI progress, while Washington and Beijing are treating it as a strategic competition that neither can afford to lose.

It is in this climate that ANSA reports, via Bloomberg, the comments of US Treasury Secretary Scott Bessent at the end of the first day of negotiations with China in New York on Sunday: "The talks were a success." Beijing and Washington have agreed to meet again on artificial intelligence. On the Chinese side, trade negotiator Li Chenggang mentions a "good atmosphere" and confirms that the discussions, which are to continue on Monday, focused on AI, trade exchanges, and investments.

For Rome, the stakes go beyond the sole Trump-Xi summit scheduled in Washington on Thursday: Gopinath hopes that the meeting will result in "at least a coordinated attempt to guarantee AI security", estimating that "the world really needs it". She links this issue to a series of successive shocks that have thickened the fog in which decision-makers are evolving: the "Liberation Day" customs duties, the war in Iran, and the energy crisis that followed, according to the image she had used as early as May 2025 at the IMF.

The Italian press currently only has second-hand reported statements on the exact content of the notification mechanism proposed by Washington; neither Bessent nor the Chinese delegation details its operation in the available excerpts. The emphasis is on the principle of a forthcoming formal dialogue, rather than its technical modalities or precise schedule.

Seoul, Monday, September 21, 2026. As Washington and Beijing announced on Sunday in New York the outline of a notification mechanism for incidents related to artificial intelligence, the South Korean press is mainly focusing on two North Korean short-range ballistic missile launches into the East Sea on the same Sunday — a sign that for Seoul, the Trump-Xi summit on Thursday in Washington is primarily being played out on the peninsula.

According to the South Korean Joint Chiefs of Staff (JCS), cited by Yonhap and the Korea Times, the two missiles were launched from the Wonsan region at approximately three-hour intervals, the first at around 15:00 and the second at 17:50. The first missile traveled approximately 450 kilometers, the second more than 600 kilometers; additional analysis was underway to clarify their characteristics. The JCS assures that South Korean and American intelligence services tracked the launches from their initial phase and shared the information with Japan.

The South Korean National Security Office (NSO) urgently convened the Ministry of Defense and the JCS after the first launch, estimating that these launches violate the UN Security Council resolutions prohibiting Pyongyang from developing its nuclear and ballistic arsenal. The Korea Times recalls that the North had already launched several short-range missiles from the same region on September 12, later presented as an artillery and missile firing exercise.

These provocations come as the summit between Donald Trump and Xi Jinping, expected on Thursday in Washington, is to determine, among other things, whether the two leaders will reaffirm their commitment to North Korean denuclearization, an objective that Pyongyang refuses to pursue by continuing its nuclear and ballistic programs, according to Yonhap. Trump has repeated his willingness to reconnect with Kim Jong-un, possibly by the end of the year, but the South Korean agency notes that it remains uncertain how far Beijing is willing to push its Cold War ally towards dialogue.

For Seoul, the announcement of an AI notification channel between Washington and Beijing adds to a list of bilateral issues — trade, rare earths, semiconductors — on which South Korea has no seat, even though its own security depends on what is said, or not, about North Korea.

Islamabad, September 21, 2026. Pakistan is following, through its economic press, the final stretch before the Washington summit between Donald Trump and Xi Jinping, scheduled for Thursday. On Sunday, September 20, 2026, in New York, US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng held an approximately eight-hour negotiation session at JPMorgan Chase's Manhattan headquarters, in the presence of US Trade Representative Jamieson Greer and Chinese international trade official Li Chenggang. Bessent described the meeting as a "very successful engagement" and specified that Washington proposed a bilateral notification mechanism for incidents related to artificial intelligence, including those affecting national security: "What we discussed was setting up a mechanism... it's going to be called the US-China AI dialogue," he stated, adding that the two delegations had agreed to meet again.

For the Pakistani specialized press, the central issue of the week remains economic. The trade truce concluded between the two powers is set to expire on November 10, and negotiators are seeking to preserve the status quo rather than announce a breakthrough: "I think status quo is probably both sides' general best expectation," notes Anna Ashton, a Sino-American trade analyst quoted by Business Recorder. The discussions also focus on Chinese flows of rare earths and critical magnets, deemed insufficient by Washington, as well as Chinese commitments inherited from the Beijing summit in May: annual purchases of $17 billion in American agricultural products and orders for over 200 Boeing aircraft.

Analyst Jon Czin, from the Brookings Institution, notes that Xi Jinping is not under pressure to make concessions and is primarily seeking to extend the tacit agreement concluded with Trump to give China time to strengthen itself. The context serves Beijing: Chinese trade has progressed over the past four months, while Trump's popularity has eroded, particularly due to the cost of the war in Iran. For Islamabad, which depends on global trade balances and critical mineral supply chains, this balance of power between the world's two largest economies weighs on regional growth prospects, more so than the technical content of the future artificial intelligence dialogue.

Singapour, 21 September 2026. Just a few days ahead of the Washington summit between Donald Trump and Xi Jinping, scheduled for Thursday, the Singaporean press is primarily focusing on the American proposal for a bilateral notification mechanism regarding artificial intelligence-related incidents. According to the Straits Times, discussions held on Sunday, 20 September, at the JPMorgan Chase headquarters in New York, brought together US Treasury Secretary Scott Bessent, Chinese Vice Premier He Lifeng, and US Trade Representative Jamieson Greer, for a full day of negotiations. Bessent presented this mechanism as a means to exchange alerts in the event of a serious incident affecting national security, stating that moving "from opacity to more transparency between the world's first and second AI powers is very important".

Channel News Asia situates the proposal within an unequal balance of power. The media outlet recalls that Chinese trade has "stagnated" since the last meeting between the two leaders in May in Beijing, while Trump's popularity has eroded, partly due to the cost of the war with Iran. Jon Czin, from the Brookings Institution, quoted by CNA, summarizes China's wait-and-see attitude: "Xi is not looking for anything concrete. He wants to extend the tacit agreement with Trump so that China has time to strengthen itself." The White House itself has sought to downplay the chances of a major breakthrough during the summit.

For Singapore, an open economy caught between the two technological blocs, the stakes go beyond the bilateral relationship alone. The mechanism mentioned by Bessent, in his own words, only covers incidents reaching "a level of national security" - leaving out the broader governance of AI that Trump and Xi had already outlined without formalizing it since May. The local press also notes that no progress was made on Sunday regarding critical minerals or export controls, despite the stated goal of extending the trade truce that expires in November. A commentary by Channel News Asia, borrowing the nuclear metaphor of the Cold War, describes the United States and China as the "two AI superpowers", recalling that their safeguards are negotiated primarily between two capitals, without third countries at the table.

Washington, September 21, 2026. As the summit between Donald Trump and Xi Jinping at the White House approaches, the US administration has put forward an unprecedented proposal: a bilateral mechanism for notifying incidents related to artificial intelligence. Treasury Secretary Scott Bessent announced this on Sunday after a day of discussions in New York with Chinese Vice Premier He Lifeng, alongside US Trade Representative Jamieson Greer. "The United States has proposed setting up a notification mechanism between our two countries," he said, specifying that it would involve exchanging alerts in the event of a serious incident affecting national security. According to Bessent, moving "from opacity to more transparency between the world's first and second AI powers is very important"; the two delegations agreed to open a formal dialogue and revisit the issue.

But Washington is strictly limiting the scope of the gesture. Greer told journalists that these AI discussions would not change US export controls targeting advanced chips and semiconductor manufacturing equipment, restrictions that continue to hinder China's technological progress. The notification mechanism therefore does not open the door to sensitive technologies; it is limited to a security incident alert channel.

Behind the scenes, the influence of the industry is weighing on the US position. Nvidia CEO Jensen Huang, described by experts as exerting "disproportionate influence" at the White House, is set to attend the state dinner organized for Xi Jinping. Trump himself has dismissed fears of uncontrolled AI, calling them a "hoax", a skepticism that analysts link to Beijing's general position. Nvidia, whose revenue soared to $215 billion in its last fiscal year compared to $17 billion in 2021, supplies the majority of the chips used by OpenAI and Anthropic.

Experts cited doubt that a substantial AI agreement will come out of the summit, citing the persistent distrust between the two capitals. The Trump administration is prioritizing voluntary safeguards focused on national security rather than binding regulation, a line that Thursday's meeting is unlikely to disrupt despite the ambitious notification channel proposal.