When US Treasury Secretary Scott Bessent spoke about artificial intelligence cooperation with China at the National Committee on US-China Relations in New York on Tuesday, he adopted a tone of guarded pragmatism, calling for building "a constructive relationship on the basis of balance and reciprocity".
Referring to the upcoming Asia-Pacific Economic Cooperation Economic Leaders' Meeting to be held in Shenzhen this month, he expressed hope that it would advance a new bilateral dialogue on artificial intelligence.
Bessent highlighted that the dialogue aims to guard against shared risks and "prevent misunderstandings between our two systems". Both countries expect to exchange best practices for evaluating and testing advanced models before their release, ensuring that human judgment remains in control of this technology, he said, adding that the two sides have agreed to establish a direct channel for alerting each other of serious incidents.
The AI dialogue between China and the US, if it proceeds smoothly, would represent a major move in global AI governance. By fostering transparency and collaboration, both nations can enhance mutual trust and mitigate potential risks, contributing to the development of the technology and the common good of humanity.
Yet the current tech blockade and restrictions imposed on China by the US have created challenges for AI collaboration and triggered concerns in US industry circles. Microsoft, along with other companies and trade groups, has reportedly urged the US administration to reconsider plans to ban imports from China of a crucial component used in AI data centers. They argue that such a ban could significantly undermine the US' competitive edge in the AI race by further delaying the rollout of data centers.
Washington has been mulling bans on imports of new models of Chinese optical transceivers, which enable data to travel over fiber-optic cables at light speed within data centers. The pretext used for such potential bans is the concern that such transceivers could be "exploited by China to steal data, install malware, or disrupt services at US data centers", claims that have never been substantiated.
As a new-generation technology, AI offers tremendous benefits but also poses potential security risks. Technological competition must not devolve into bloc confrontation. Countries should focus on building collaborative frameworks that emphasize common ethical standards and transparent practices, foster innovation and safeguard against misuse.
As major countries in AI development, China and the US have differences in their systems, technological paths and models. Yet they share a common responsibility to mitigate risks and prevent the misuse of the technology, which provides a foundation for cooperation. The willingness to communicate on AI cooperation, as expressed in Bessent's remarks, is encouraging. However, policies should not remain mere rhetoric. Advocating cooperation on AI safety measures while planning to disrupt the supply chains for key components will only erode mutual trust and make dialogue more challenging.
China and the US should prioritize cooperation while keeping competition within reasonable guardrails. There is no reason for them not to engage in healthy competition that can spur innovation, allowing them to leverage their respective strengths to drive technological advancement and effectively manage the risks. Fragmenting global supply chains and politicizing commercial and technological issues will ultimately harm the entire AI ecosystem and prevent humanity benefiting from its development.
Multilateral platforms such as APEC can serve as venues for bridging gaps and building consensus on AI governance. The APEC Economic Leaders' Meeting should be used to bring all parties to the table around practical plans for model testing as well as risk management and control. A collaborative move by the participants at the APEC meeting could set a precedent for international cooperation in AI, highlighting the importance of coordinated efforts to address global challenges and ensure the responsible development of this transformative technology.
AI knows no national boundaries, and its risks are trans-boundary. It would be helpful for Washington to move beyond zero-sum thinking, respect market rules and heed the voices of the business sector. That can help China and the US better manage differences through communication, address common risks through cooperation and jointly advance global AI governance amid healthy competition for the benefit of humanity.
