In August, Chinese Foreign Minister Wang Yi met with U.S. National Security Adviser Jake Sullivan, and both sides agreed to hold the second round of the U.S.-China Governmental Dialogue on Artificial Intelligence and other institutional arrangements at an appropriate time.
Three months ago, China and the United States initiated their first round of governmental dialogues on AI, a good beginning for bilateral cooperation in promoting “AI for Good” between the two leading actors in AI. The dialogues demonstrate that both China and the U.S. share common interests in AI, though their bilateral ties are dominated by fierce strategic competition.
Actually, there is room for China and the U.S. to cooperate on AI, as both sides face the global challenges it poses, such as privacy protection, data security, and ethical issues. Nonetheless, cooperation still faces uncertainties due to fluctuations in relations. In fact, they have not yet formulated a regular mechanism on AI cooperation and only recently had their first round of dialogue on the subject.
What’s more, deep mutual distrust is a major barrier for AI cooperation between China and the United States. Although the U.S. initiated the AI dialogue with China, it has not abandoned its containment policy on China’s AI development. So it is reasonable for China to be suspicious about America’ true intentions on AI dialogue.
Given the uncertainties and mistrust on AI cooperation, both governments should demonstrate a firm commitment to holding long-term intergovernmental dialogues. Both sides should contemplate instituting a long-term mechanism that minimizes disruptions, suspensions or terminations arising from a deterioration in bilateral relations.
They should also consider expanding dialogue channels to increase the certainty and continuity of cooperation. Apart from the existing governmental dialogue organized by China’s Department of North American and Oceanian Affairs of the Ministry of Foreign Affairs and the corresponding U.S. agencies within the Department of State and the White House, a 1.5-track dialogue mechanism could be established at a subordinate level.
Led by the deputy director of China’s Department of North American and Oceanian Affairs and U.S. counterparts, this mechanism could involve staff from other relevant Chinese government agencies such as the Ministry of Science and Technology, Ministry of Industry and Information Technology, Ministry of Commerce and the Cyberspace Administration of China, as well as university professors and think tank experts, senior managers and engineers from the top AI companies of both countries. Multiple discussion groups could be formed based on different topics to facilitate specific in-depth discussions and propose practical solutions.
Beyond maintaining intergovernmental dialogues, non-governmental AI exchanges between China and the U.S. should also be enhanced. Facilitating non-governmental channels could broaden cooperation opportunities and improve public perceptions of AI in both countries.
The report on the 2-track dialogue released in April by Tsinghua University’s Center for International Security and Strategy showed that both sides have achieved consensus and progress in four areas: identifying the potential security risks of AI, clarifying AI terms, analyzing conflict mechanisms and controlling measures related to the military use of AI and exploring the path of AI security governance.
Chinese universities and think tanks, particularly those with expertise in AI, can collaborate with relevant U.S. institutions. Leading AI enterprises in China, including Baidu, Alibaba and Tencent, can also attempt AI exchanges with top U.S. companies.
So far, China and the U.S. have incorporated security governance and ethical norms on AI into their institutional frameworks through laws, regulations, and policy documents, aiming to establish a globally coordinated and efficient AI governance system. In that case, China and the U.S. might want to consider forming a similar mechanism like the U.S.-EU Trade and Technology Council — in which both sides have achieved phased consensus on the core issues on AI governance — to deepen cooperation on that matter.
The potential cooperative areas mainly include AI system auditing standards, joint testing mechanisms and shared indicator databases for credibility and risk management methodologies. These low-sensitivity areas might provide a breakthrough for cooperation.
Nevertheless, one should not underestimate the difficulties involved in China-U.S. cooperation on AI governance. Beyond the differences in their respective laws and regulations on core issues, including the ownership of patents, the definition of copyright and infringement liability, America’s “pan-securitization” on cooperation in all fields is yet another main obstacle for AI cooperation. Both sides should focus more on the ethnic issues of AI governance at the early stage of bilateral AI cooperation
In short, AI cooperation has emerged as an area with high potential for China-U.S. relations. It would have a lasting impact. Therefore, both sides should take more effective measures to promote it to generate positive energy for bilateral ties and future human development.
