Today, the head of the Department of American and Oceanian Affairs at China’s Ministry of Commerce provided a public briefing on the outcomes of the eighth round of U.S.-China economic and trade consultations.
MOFCOM Official Explains Outcomes of the Eighth Round of China-U.S. Economic and Trade Consultations
From September 20 to 23, the Chinese and U.S. economic and trade teams held consultations in New York and Washington, D.C., reaching a number of positive understandings and contributing economic and trade outcomes to the meeting between the two heads of state. An official from the Department of American and Oceanian Affairs of China’s Ministry of Commerce provided the following explanation of the outcomes of the consultations.
I. The China-U.S. Board of Trade
Following consultations, China and the United States agreed to establish a China-U.S. Board of Trade under the China-U.S. economic and trade consultation mechanism, with the primary task of optimizing bilateral trade. The two economic and trade teams have reached agreement on the Board’s mandate, structure, responsibilities and consultation arrangements, which will be announced at a later date.
The Board of Trade will begin by discussing the “$30 billion for $30 billion” reciprocal tariff reduction framework, with a view to reaching agreement. An agricultural working group has been established under the Board, which will also consider other arrangements for optimizing bilateral trade.
It is believed that the Board of Trade will provide an important platform and institutional support for China and the United States to continuously lengthen the list of cooperation and shorten the list of problems in the economic and trade field, and promote the sustained, stable and positive development of China-U.S. economic and trade relations.
II. The “$30 Billion for $30 Billion” Reciprocal Tariff Reduction Arrangement
To implement the important understandings reached by the two heads of state, China and the United States, on the basis of mutual respect and consultations on an equal footing, held multiple rounds of discussions under the Board of Trade framework on the $30 billion reciprocal tariff reduction arrangement and have now reached agreement.
The two sides agreed to reciprocally reduce tariffs on approximately $30 billion worth of products imported from each other. Tariffs on approximately 90 percent of the products on each side’s list will be reduced to most-favored-nation (MFN) tariff rates.
After completing the relevant procedures in accordance with their respective domestic laws, the two sides will implement the tariff reduction arrangement simultaneously. Further details of the reciprocal tariff reduction arrangement will be announced at a later date.
This arrangement will help further stabilize China-U.S. economic and trade relations and create more favorable conditions for exports of relevant Chinese products to the United States.
III. Expanding Coal Trade
The Chinese and U.S. economic and trade teams reached an understanding that tariffs on China’s imports of U.S. coal will be included in the “$30 billion for $30 billion” reciprocal tariff reduction framework. This will facilitate China’s imports of U.S. coal in both 2027 and 2028.
Imports of U.S. coal will provide a useful supplement to China’s domestic coal market while generating stable economic benefits and employment for the U.S. coal industry, thereby helping to stabilize and expand bilateral trade between China and the United States.
It is hoped that, through the joint efforts of both sides, China-U.S. cooperation in the coal sector will continue to deepen and deliver mutually beneficial and win-win outcomes.
IV. The China-U.S. Investment Board
Following consultations, China and the United States agreed to establish a China-U.S. Investment Board under the China-U.S. economic and trade consultation mechanism.
The Board will provide an institutionalized platform for communication between the two sides. The Chinese and U.S. economic and trade teams will maintain regular dialogue on potential investment opportunities and investment barriers and, subject to their respective laws and regulatory requirements, enhance policy transparency and predictability and address the legitimate concerns of businesses.
V. Agricultural Working Group
In May this year, the Chinese and U.S. economic and trade teams reached a series of positive understandings on promoting the resolution of certain non-tariff barriers and market-access issues affecting agricultural products on both sides. They also agreed in principle to include the relevant products in the “$30 billion for $30 billion” reciprocal tariff reduction framework.
The two teams subsequently held professional and pragmatic consultations on agricultural issues. Against this backdrop, the two sides agreed to establish an agricultural working group under the China-U.S. Board of Trade, co-chaired by China’s Ministry of Commerce and the Office of the U.S. Trade Representative, with the participation of relevant regulatory authorities from both countries.
The working group will discuss two-way market access for agricultural products, regulatory issues and other related matters. The two sides agreed to hold the first meeting of the working group before the end of 2026.
China will continue to use the working group meetings to urge the U.S. side to address China’s concerns in the agricultural sector.
VI. Financial Services
In recent years, China has independently expanded the opening-up of its financial services sector, which has received recognition from the international community and benefited many U.S.-funded institutions.
During the latest consultations, the two sides reached an understanding in principle on financial services. China will review and approve, in accordance with laws and regulations, applications by financial services institutions from the United States and other countries to conduct business and establish branches in China.
China hopes that the United States will likewise provide Chinese-funded financial institutions with a fair, transparent and stable policy environment.
VII. China-U.S. AI Dialogue
During this round of consultations, the two sides agreed to establish an AI dialogue under the China-U.S. economic and trade consultation mechanism, with Vice Premier He Lifeng and Treasury Secretary Scott Bessent serving as the respective Chinese and U.S. principals.
The two sides held the first dialogue and exchanged views on the risks and benefits associated with AI. They agreed to hold the next dialogue before the end of November this year.
The two sides also agreed to establish a communication channel for AI incidents.
VIII. Extension of the Joint Arrangement Reached in the Kuala Lumpur Economic and Trade Consultations
In October 2025, China and the United States reached a joint arrangement in Kuala Lumpur to address their respective economic and trade concerns, under which each side suspended the implementation of certain tariff and non-tariff measures until November 10, 2026.
During the latest consultations, China and the United States agreed that, on the basis of extending the joint arrangement reached in the Kuala Lumpur consultations until January 10, 2027, they would continue to actively explore the issue with a view to reaching a solution agreed upon by both sides.
This arrangement, on the one hand, provides space for both sides to take stock of and assess the implementation of the joint arrangement and consider how to move China-U.S. economic and trade relations forward. On the other hand, by combining an extension for a defined period with continued active discussions, it provides a relatively stable and predictable policy environment for cooperation between Chinese and U.S. businesses.
It is believed that the two sides will continue their efforts and, through high-level economic and trade interactions before the end of the year, seek a positive solution regarding a further extension.
IX. Increasing Direct Passenger Flights Between China and the United States
Passenger flights between the two countries play an important role in promoting people-to-people exchanges, cultural and social interactions, and economic and trade cooperation.
During the latest consultations, China and the United States exchanged views on increasing direct passenger flights. The two sides agreed to remain in communication regarding an increase in China-U.S. flights and related matters.
X. Implementation of Understandings Reached in Previous Consultations
Since last year, China and the United States have held eight rounds of economic and trade consultations and reached a series of mutually beneficial, balanced and positive economic and trade outcomes on important issues of common concern, including the promotion of bilateral trade and investment. These outcomes have helped maintain the overall stability of China-U.S. economic and trade relations.
During the latest consultations, China and the United States reaffirmed and once again committed to continuing to implement the understandings reached in previous rounds of economic and trade consultations.
This will create favorable conditions and a stable environment for economic and trade cooperation between the two countries and will also contribute to global economic development.
The Office of the U.S. Trade Representative has published lists of non-strategic U.S. and Chinese goods for which the two sides will consider lowering tariffs, as well as the terms of reference and working procedures for the U.S.-China Board of Trade, which will serve as the forum for those discussions.
On the reciprocal “$30 billion for $30 billion” tariff reduction arrangement, the U.S. side has emphasized that the products involved are “non-sensitive goods,” while the Chinese side has stressed the idea of “lengthening the list of cooperation and shortening the list of problems.” These are largely differences in emphasis; the basic structure described by the two sides is consistent. China has selected a group of products it imports from the United States with an annual trade value of roughly $30 billion based on 2024 figures, while the United States has selected a group of Chinese products worth approximately the same amount. The U.S. side first published China’s import list, saying that roughly 30 percent of U.S. exports to China fall within the scope of the arrangement, including agricultural products, seafood, cosmetics and medical devices.
For now, however, what has been finalized is the product lists rather than actual tariff reductions. Judging from the U.S. releases, the arrangement appears to work roughly as follows: each side first identifies approximately $30 billion worth of products and places them in a pool of goods eligible for potential preferential treatment. Future tariff reductions would focus first on these products, with the specific reductions to be agreed upon later and then implemented through each side’s domestic procedures. U.S. Trade Representative Jamieson Greer made a similar point in a media interview on Sunday, saying that this did not mean tariffs on these goods would be cut immediately. Rather, the two sides were setting these products aside so that they could be treated separately when future tariff adjustments or other trade measures were considered.
The Chinese briefing, however, suggests that the tariff reduction package has already been agreed at the political and negotiating levels, with only the respective domestic implementation procedures remaining before the two sides implement the reductions simultaneously. China also disclosed one important detail: of the approximately $30 billion worth of products on each side’s list, tariffs on around 90 percent of the products will be reduced to most-favored-nation (MFN) rates.
The U.S. side also published the terms of reference for the 30-for-30 reciprocal tariff framework. Overall, the U.S. description again reads more like this: the two sides first identify approximately $30 billion worth of products each and place them in a pool of goods eligible for potential preferential treatment. Future tariff reductions would focus on these products, with the specific treatment to be negotiated and then implemented through each side’s domestic procedures. As Greer put it in an interview with Fox News on Sunday: “These are not immediate tariff cuts or things like this; these are goods where if there are trade measures in the future or things that we’re looking at, we just leave these off to the side.” The two sides do not envision adjusting the reciprocal tariff lists more frequently than once a year.
There are at least two possible explanations for the difference in wording. One is that the two sides have in fact already agreed on the substantive tariff package, but because the United States still needs to complete its domestic legal procedures, the formal U.S. documents use more cautious language. The other is that the two sides have reached political agreement on the overall framework, while some of the specific tariff rates still need to be finalized through U.S. domestic procedures.
On the agricultural working group and the Investment Board, the two sides’ descriptions are also broadly consistent, with no obvious readout gap. China said that an agricultural working group has already been established under the Board of Trade, co-chaired by China’s Ministry of Commerce and USTR, to discuss two-way agricultural market access and regulatory issues. The first meeting is planned before the end of 2026. The White House had already confirmed that the Board of Trade had launched a working group focused on agricultural market-access barriers, while the working procedures released on September 27 explicitly provide for the establishment of an agricultural working group. Reuters also reported on September 28 that the first meeting of the working group would take place before the end of the year.
The descriptions of the Investment Board are even more closely aligned. China said that the two sides would maintain regular dialogue on “potential investment opportunities and investment barriers” and establish an institutionalized communication channel. The White House similarly described the Board as a forum for discussing “potential investment opportunities and investment-related impediments” and as a “structured channel” for addressing investment issues of practical commercial significance.
On coal, the U.S. side has disclosed more specific details than the Chinese side. China confirmed that U.S. coal would be included in the “$30 billion for $30 billion” framework and said that this “will facilitate” China’s imports of U.S. coal in 2027 and 2028. The Chinese announcement, however, did not specify an annual purchase volume. The White House was much more specific: China will import at least 10 million metric tons of U.S. coal in 2027 and at least another 10 million metric tons in 2028. Reuters also described this on September 28 as a coal-import arrangement reached at the summit.
On the extension of the joint arrangement reached in the Kuala Lumpur economic and trade consultations, China confirmed the U.S. announcement that the arrangement had been extended until January 10, 2027. At the same time, however, China emphasized that the current arrangement is only an interim arrangement—or, in the Chinese side’s own words, a “foundation.” The two sides will continue actively exploring the issue with the aim of reaching a mutually agreed solution. The implication is that China does not regard the current two-month extension as a satisfactory or final outcome. China also said it believed the two sides would continue working through high-level economic and trade interactions before the end of the year to find a positive solution on a “further extension.”
The Chinese briefing also addressed the U.S.-China AI dialogue:
During this round of consultations, the two sides agreed to establish an AI dialogue under the U.S.-China economic and trade consultation mechanism, with Vice Premier He Lifeng and Treasury Secretary Scott Bessent serving as the respective Chinese and U.S. principals. The two sides held the first dialogue and exchanged views on the risks and benefits associated with artificial intelligence. They agreed to hold the next dialogue before the end of November this year. The two sides also agreed to establish a communication channel for artificial intelligence incidents.
At the Foreign Ministry’s regular press conference on the same day, the spokesperson said that maintaining communication and exchanges with the United States through channels including the intergovernmental dialogue on AI is part of what it means to build a constructive U.S.-China relationship of strategic stability based on respect, fairness and reciprocity.
On September 28, Foreign Ministry spokesperson Guo Jiakun chaired the regular press conference.
A foreign media reporter asked: One of the outcomes of the U.S.-China summit was an agreement to establish an AI dialogue mechanism, including a communication channel for AI-related incidents. Before China recently established an international organization for AI cooperation, China had consistently supported multilateral cooperation on artificial intelligence and advocated a more comprehensive approach, including the development of ethical guidelines and relevant procedures to ensure that AI remains under human control. The U.S. position, by contrast, appears to favor a more limited approach focused on AI safety and warning mechanisms between specific parties. At present, the new U.S.-China mechanism appears to reflect the U.S. approach to a greater extent. Is China adjusting its position on international AI governance? And is this bilateral mechanism sufficient to constitute a framework for global AI governance?
Guo Jiakun said that China has consistently participated in global AI governance in a constructive and responsible manner and has continued to contribute Chinese proposals. China has always maintained that the United Nations should play the central role in building a fair and equitable global AI governance system.
“At the same time, we are also willing to maintain communication and exchanges with the United States through channels including the intergovernmental dialogue on AI. This is both an important pathway for promoting the development of AI for good and for the benefit of all, and an integral part of building a constructive U.S.-China relationship of strategic stability based on respect, fairness and reciprocity,” Guo said.
President Trump said he had discussed with President Xi the idea of replacing the term “artificial intelligence” with “super intelligence,” and indicated that President Xi also liked the term “super intelligence.” On September 26, however, the Chinese Foreign Ministry specifically responded that China respects the U.S. preference for the term “super intelligence” but will continue to refer to it as “artificial intelligence.” China did not announce any agreement to adopt a common name.
That said, Liu Chong, Director of the Center for Science, Technology and Strategy at the China Institutes of Contemporary International Relations (CICIR), has already publicly used the term “super intelligence” in an article published on the institute’s official WeChat account and offered a substantive discussion of the “governance of super intelligence.”
Full translation of Liu’s article is below:
Superintelligence Governance Should Be a Systemic Undertaking Built on Mutual Trust and Cooperation
超级智能治理应当是互信合作的体系性工程
Author: Liu Chong
Institution: China Institutes of Contemporary International Relations
Publication date: September 28, 2026
Superintelligence governance is a major issue concerning the future and destiny of humanity. Dialogue between China and the United States is an important step for the two countries to shoulder their responsibilities as major powers and safeguard humanity’s future.
Liu Chong
Director, Center for Science, Technology and StrategyFrom September 23 to 25, 2026, Chinese President Xi Jinping paid a state visit to the United States. During the visit, the two heads of state reached consensus on eight outcomes, including the following: “The two sides agreed to establish a China-U.S. dialogue on artificial intelligence to exchange views on AI-related risks and benefits. The next dialogue will be held in November this year. The two sides agreed to establish a communication channel for artificial intelligence incidents.” The White House outcome document specifically stated that the two leaders agreed to use “super intelligence,” rather than “artificial intelligence,” to refer to the emerging technology covered by these discussions.
As the two countries currently at the forefront of artificial intelligence development, China and the United States have reached an important arrangement that represents an entirely new governance scenario and model in human history: governance must begin even though its object has not yet fully emerged and does not even have a universally accepted definition. Such governance is exceptionally important because of AI’s unprecedented impact, yet highly unstable because of the rapid iteration of the technology. This poses a new test for China, the United States, and the entire world. It also means that governance must be a systemic undertaking that continuously identifies risks, anticipates trends, makes forward-looking adjustments, shares benefits, and builds trust as the technology evolves.
Superintelligence Risk Is Not a “Red Line,” but a Shifting “Range”
People often imagine the development of artificial intelligence as a clearly defined ladder: narrow AI, artificial general intelligence, and then, after crossing a certain threshold, the era of superintelligence. Yet it has become a broad industry consensus that no clear red line can be drawn, with safety and control below it and “The Matrix” above it. Reality also amply demonstrates that risks will not wait until the label “superintelligence” formally appears before suddenly materializing. In recent months, frontier AI companies including OpenAI and Anthropic have encountered dangerous instances in which models in training independently gained unauthorized access, prompting these companies on several occasions to suspend the relevant model training. These incidents show that the real-world “braking point” does not lie in whether a model has formally been classified as AGI or superintelligence. Increasingly, it depends on whether existing control systems can still reliably constrain the model’s capabilities and behavior.
Rather than futilely searching for an abstract “superintelligence red line,” a more practical direction for governance may be to establish a dynamic “risk range–response level” system: which combinations of capabilities signal a marked rise in risk; which anomalous behaviors require intensified monitoring; under what circumstances training should be slowed or even suspended; under what circumstances it may resume after corrective action; and which behaviors indicate that a particular training pathway should be terminated.
“Shared Understanding” Is a Prerequisite for “Effective Joint Governance”
If the boundaries of superintelligence risk are themselves changing rapidly, then the first public good that international governance needs to provide is not a strict and exhaustive set of rules, but shared knowledge. This is precisely where China and the United States first need to align their understanding.
China and the United States do not need to adopt identical AI industrial policies, nor do they need to share model parameters or core technologies. They do, however, need to gradually develop a risk vocabulary that both sides can understand: What constitutes dangerous autonomy? What degree of “model overreach” should be elevated to the level of a security incident? How should major AI risks and accidents be classified? Which training pathways are more likely to carry risks that outweigh their benefits? … The safety lessons now scattered within a small number of companies should gradually be transformed into a common language that can be understood and discussed across companies, research institutions, and countries.
U.S. companies remain ahead in model research and development, and the models they develop and train internally are more advanced than those available for commercial use. Yet all U.S. frontier models are closed-source. This “black box” characteristic has in fact heightened external concern and anxiety. U.S. companies should set an example by strengthening the disclosure of risk information.
In addition, superintelligence risk governance cannot allow a small “inner circle” of frontier companies to determine the “overall future” of humanity. The Chinese and U.S. governments should engage in full exchanges on how to strengthen corporate oversight and coordinate development and security. Through dialogue, they should foster an interdisciplinary and cross-sector community spanning law, artificial intelligence, cybersecurity, strategic studies, and public policy, thereby steadily building the capabilities needed for future global governance.
Both Development and Security Should Be Achieved Together
The China-U.S. dialogue on artificial intelligence is intended to “exchange views on AI-related risks and benefits.” An especially important part of this should be how China and the United States can provide public goods. Artificial intelligence is, first and foremost, a new frontier of human development. If superintelligence governance focuses only on restricting proliferation, erecting barriers, and raising thresholds, it will be difficult to build a genuinely broad international foundation. If some countries develop while others do not, and some countries possess capabilities while others do not, this may become the greatest form of global inequality in the age of intelligence.
In the field of development, China and the United States should engage in extensive exchanges, explore cooperation, and identify institutional arrangements that can turn technological dividends into opportunities for global development. Development gains should benefit all humanity more broadly and equitably, ensuring that no country and no individual is structurally excluded by technological progress from living a life of dignity. The ultimate purpose of AI development is not to create more advanced machines, but to create better lives; it is not to make intelligence a privilege enjoyed by a few people or a few countries, but to enable all humanity to share the development opportunities brought by technological progress.
In the field of security, the dangers are self-evident if highly automated cyberattacks, biological design capabilities, and other high-risk capabilities can be replicated at low cost. Yet the long-term concentration of advanced intelligence in a small number of countries and companies may cause even greater harm. If a few countries possess advanced systems that use intelligent technologies for cyber offense and defense, while many small and medium-sized countries still rely mainly on traditional defensive technologies, the AI era will create not merely a “digital divide,” but potentially a new “security divide.” Mature, secure, and broadly applicable intelligent capabilities should, wherever possible, be transformed into widely available public capabilities. In other words, superintelligence governance should not only prevent the dangerous proliferation of the “spear,” but also promote the spread of the “shield.” Global cybersecurity capacity-building, the sharing of AI safety tools, the development of intelligent defense systems for small and medium-sized countries, and international mechanisms for risk early warning and coordinated response should all be gradually incorporated into the governance agenda.
The Continuity of Human Civilization Requires Joint Stewardship
Terence Tao and 24 other Fields Medalists recently issued a joint open letter noting that AI companies are increasingly treating the “solution of major mathematical problems” itself as a benchmark of model capability, potentially detracting from the truly important task of “forming new conceptual understanding and insights.”
If AI’s production of cross-domain knowledge and its pace of problem-solving far exceed humanity’s ability to understand and absorb knowledge, it may greatly increase the productivity of the best scientists of this generation while weakening the training system on which the growth of the next generation of outstanding scientists depends. The modern scientific community consists of more than a tiny number of people at its pinnacle. A leading researcher typically passes through a long progression as a student, graduate student, and rank-and-file researcher, gradually developing professional judgment and creativity through problem-solving, reading, experimentation, failure, and reflection. If AI largely replaces these processes, how will researchers be trained and selected? And in the future, who will determine whether the proofs, programs, and scientific conclusions produced by AI are genuinely trustworthy?
Superintelligence governance therefore cannot discuss model safety alone. It must also address the broader issue of civilizational security by preserving the systems through which humanity understands and produces knowledge and transmits the capacity for innovation from one generation to the next.
Continuously Expanding the Space for Trust amid Competition
All of these questions ultimately return to a practical contradiction: the opportunities and risks associated with superintelligence create a clear need for cooperation, while AI competition between China and the United States continues to intensify.
This tension is already evident in the public positions of some U.S. frontier AI companies. Anthropic has explicitly argued that the United States and its allies should maintain their lead in the most advanced areas of AI, and has advocated further restrictions on China’s access to advanced chips, model capabilities, and related technologies to ensure that so-called “democratic countries” retain leadership over AI development. Such arguments and restrictive policies inevitably make China wonder whether so-called governance cooperation is an institutional shackle designed to constrain its development opportunities. Conversely, the United States will likewise worry that exchanges of risk information could evolve into exchanges of capability information, and that security cooperation could help a competitor narrow the technological gap.
A classic security dilemma may then emerge: each side, fearing that the other will achieve a breakthrough first, accelerates its own research and development while restricting the other; those restrictions further heighten the other side’s sense of insecurity; insecurity makes risk assessment and security cooperation more difficult; and the more difficult security cooperation becomes, the more reason both sides have to accelerate competition further.
The greatest challenge for China-U.S. superintelligence governance is therefore whether the two sides can preserve sufficient space for trust amid strategic competition and, on that basis, steadily expand the cooperation that is necessary. The goal of future dialogue should be to establish a baseline risk-cooperation framework that will not easily fail even if technological competition between China and the United States persists or intensifies. Beginning with the most basic alignment of risk assessments through a dialogue on superintelligence, the two sides should gradually develop a shared risk vocabulary, the scientific trust that is necessary, methods of capability-sharing that balance security and development, and a strategic consensus that allows them to apply the brakes when necessary.
Superintelligence governance is a major issue concerning the future and destiny of humanity. Dialogue between China and the United States is an important step for the two countries to shoulder their responsibilities as major powers and safeguard humanity’s future. For China and the United States, the real test is not whether they can stop competing, but whether they can build mutual trust and sustain cooperation amid competition. Even as they pursue competitive advantage, they must align their understanding of risks, manage strategic risks, advance global development, and carry human civilization forward, thereby making a decisive contribution to the systemic undertaking of global AI governance.