During the November 2023 summit in San Francisco, Chinese and American leaders reached an agreement to establish a intergovernmental dialogue on artificial intelligence, marking the beginning of a new chapter in China-U.S. cooperation on AI.
In January, Wang Yi, a member of the Political Bureau of the Communist Party of China Central Committee, held a new round of talks with U.S. National Security Adviser Jake Sullivan in Bangkok. The two sides agreed to hold their first intergovernmental dialogue on AI this spring, indicating the increasing importance of the topic in the bilateral relationship.
Currently, both China and the United States possess rich talent resources, robust economic capabilities and cutting-edge computing power in the development of complex AI models. The interactions of the two countries in the emerging field, as well as in global governance, will have a significant impact on the future of humanity. From a strategic perspective, as leaders in the development of AI technology China and the U.S. have a responsibility to cooperate, manage differences and avoid risks. They not only need to avoid the risk of uncontrolled proliferation of AI technology but also need to prevent a slide into the trap of a zero-sum game because of a lack of mutual trust.
From a technical standpoint, China and the U.S. have the capability to collaborate in the research and governance of AI, as development of its in both countries is highly complementary. With respect to the three core elements — data, computing power and algorithms — China has an advantage in data, while the U.S. is more powerful in terms of computing power and algorithms. In the application domain, the countries emphasize different things, and so their cooperation would be beneficial in achieving resource complementarity and in promoting technological progress.
China and the U.S. also have a certain consensus on the basic principles, ethics, standards, risk assessment and international cooperation needed for AI development. First, both countries emphasize principles of responsibility, security and reliability in the AI technology. Second, they prioritize privacy protection and source transparency in ethical standards, striving to avoid bias and discrimination in data and algorithms. Third, both countries are committed to advancing the development of unified technical standards and frameworks, as well as to the construction of an AI risk-assessment system to enhance the interpretation and predictability of AI. Fourth, they both intend to expand international cooperation and dialogue, working with multiple parties to formulate principled and coordinated programs for international cooperation.
Despite having achieved the basic conditions for cooperation in terms of intent and capability, difference remain between China and the U.S. in development concepts, values, bundling and legal aspects. These differences need to be addressed through enhanced communication in future Track One and Track Two dialogues as soon as possible. This will lay the foundation for cooperation. First, China suggests upholding a people-centered approach in AI development, while the U.S. only requires ensuring human involvement in decision-making as it calls for “responsible AI.”
Second, China and the U.S. have different cultural traditions, resulting in varying interpretations of philosophical concepts and different priorities. Third, there are differences between the two countries on issues such as intellectual property, privacy protection, technical standards, technical norms and regulatory approaches and strengths, making it challenging to reach a coordinated consensus in the short term.
Further, influenced by its so-called strategic competition with China, the U.S. advocates merging values with technological leadership in its AI strategy. The U.S. aims to advance its leadership in both technology and standards in the global AI field and, under the influence of countering China, is pressing China through the formation of a
technology alliance. In the short term, the U.S. policy of “small yard, high fence” in the technology sector is difficult to reverse. Finding a balance between hardware and regulatory cooperation amid technological pressure poses a serious challenge for both countries.
Unlike issues with a stronger security element, such as nuclear technology, and topics with a more apparent developmental aspect, such as carbon emissions, artificial intelligence simultaneously possesses strong challenges in both. Thus we still need to ask a further question: What role should this new technology play in U.S.-China relations?
Regardless what the answer is, both sides should adopt a more positive attitude, engage in practical research and multifaceted communication and correctly recognize and evaluate each other’s goals and core concerns. They should clarify the areas of consensus and difference in the AI field, gradually solidifying the foundation for cooperation and trust and dispelling doubts. China and the U.S. should not only actively promote the healthy and orderly development of AI in both countries, but also make meaningful contributions to global AI governance.
With respect to dialogue, the two countries have established channels for multilateral and bilateral Track Two dialogues in the field of artificial intelligence, laying an important foundation for direct intergovernmental dialogue. On the multilateral front, since 2019, China, the U.S. and other countries, have engaged in discussions on the principles of AI development, military applications and ethical norms. This was done within the frameworks of intergovernmental international organizations such as the United Nations and the G20. They also jointly signed the Bletchley Declaration at the first AI Safety Summit in the UK last year.
On the Track Two front, Tsinghua University’s Center for International Security and Strategy and the Brookings Institution in the U.S. have been conducting dialogue since 2019. This project is one of the earlier initiatives for Track Two dialogues in this field. The mechanism allows both sides to engage in candid discussions as they address differences, enhance mutual understanding and explore potential areas for future cooperation.
In the future, both China and the U.S. should consider moving forward in the following three areas:
First, they should continue to support and advance multilateral dialogue. Within the framework of the United Nations and in light of new developments, they should jointly support the work of the UN High-Level Advisory Body on Artificial Intelligence and should supplement and revise global agreements, such as the Convention on Certain Conventional Weapons and the UNESCO Recommendation on the Ethics of Artificial Intelligence. They should also promote the construction of international governance mechanisms within frameworks such as regional AI safety and security summits.
Second, China and the U.S. should continue to encourage and support various channels of civil dialogue. Through various mechanisms such as think tank dialogues and business collaborations, both can engage in more in-depth exchanges around key issues in a flexible and candid manner. This has the potential to enhance understanding of each other’s positions, explore priorities for AI cooperation and identify paths for addressing differences. This could be a beneficial supplement to intergovernmental dialogues.
Additionally, given the significant roles of companies in both China and the U.S. in the field of AI technology, there should be increased interaction and engagement between relevant enterprises in both countries on technology research and development, industrial policies, industry standards, data management and social responsibility. Companies should be encouraged to innovate actively, identify practical problems and opportunities and participate together in AI governance.
Third, the two countries should seize the opportunities presented by the stabilization and improvement of bilateral relations after the San Francisco summit. They should promptly discuss the structure of the intergovernmental dialogue mechanism, agenda-setting and the selection of negotiators. In the initial stages of dialogue, the two sides can start by exchanging views on AI-related terminology, policy propositions, talent development and global governance.
They can also engage in exchanges regarding the fundamental principles of AI development, ethical concepts and related terminology and their respective domestic policy initiatives to bridge cognitive gaps. They can also enhance the exchanges of AI-related professionals, encouraging students, scholars and experts to learn from each other, engage in joint research projects and collaborate on various fronts.
And with a focus on global risks and challenges, both countries should explore the application of AI technology in areas such as climate governance, drug control, crime prevention and counter-terrorism, with a view toward cooperation and making greater contributions to global governance.