Sino-US cooperation on AI on the table

Artificial intelligence could emerge as an area of bilateral cooperation between the United States and China, experts said, following a summit between the two countries' leaders in Beijing that spurred discussions on AI.

During US President Donald Trump's recent visit to China in May, both heads of state had constructive exchanges on artificial intelligence and agreed to hold dialogue between the two governments on this issue.

Chinese Foreign Ministry spokesperson Guo Jiakun said at a news briefing on May 19 that the two nations, as leading AI powers, need to work together to promote the development and improve the governance of AI to make sure that it will better contribute to the progress of human civilization and the common welfare of the international community.

Fred Teng, president of the America China Public Affairs Institute and a fellow of the Foreign Policy Association, told China Daily that AI can no longer be treated as a purely technical matter.

"AI is now an economic, security, governance and strategic stability issue," said Teng, stressing that it's important for the two countries to collaborate on addressing risks posed by malicious use of AI.

He stressed the urgency of acting now, warning that AI capabilities are advancing faster than diplomacy, regulation or public understanding, and that tools once exclusive to governments may soon be available to smaller groups or individuals.

"AI can make cyberattacks more automated, deepfakes more convincing, fraud more scalable, and biological or chemical risks harder to control. These dangers do not respect borders," said Teng. "AI is not separate from strategic stability. AI is now part of strategic stability."

Zhiqun Zhu, professor of political science and international relations and director of the China Institute at Bucknell University, echoed the urgency of collaboration to prevent such dangers.

"AI technologies are progressing rapidly and changing the world enormously, so there is an urgent need to establish some protocol about the usage of AI at the international level. The United States and China are the two biggest technological powers and should take the leadership in developing such protocols," he told China Daily.

Zhu acknowledged, however, that translating shared interests into concrete cooperation would be difficult. He said that agreeing on where and how AI can be used, particularly in addressing security challenges, would be "a tall order".

Uniquely positioned

However, Kyle Chan, a fellow in the John L. Thornton China Center at the Brookings Institution, argued that as developers of the world's most advanced AI systems, the two countries are uniquely positioned to address these risks, and that doing so does not require broad trust, strategic alignment or compromise on national interests.

"The United States and China can continue to compete vigorously in AI while taking practical steps to reduce shared risks," Chan wrote in a recent commentary published by the Brookings Institution.

For cooperation, the two countries can start with several practical areas, said Teng, who also serves as an adviser to the George H. W. Bush Foundation for US-China Relations.

He called on both governments to align AI safety standards for advanced models that could affect cyber operations, biological research, financial systems, critical infrastructure or military decision-making.

The two sides should also build risk-reduction mechanisms, including emergency communication channels, for use in the event of an AI-enabled attack on critical systems, said Teng. Information-sharing on malicious uses of AI, from cybercrime and deepfakes to threats against hospitals and power grids, should also be on the table, he added.

When it comes to the military domain, Teng said both governments need to agree on foundational principles — above all, that human beings must remain responsible for life-and-death decisions, even as AI systems grow more capable.

He drew a parallel to Cold Warera crisis management, noting that Washington and Moscow cooperated on nuclear risk reduction even as they competed fiercely. The same logic, he said, applies to AI today.

"Cooperation can be limited to areas of shared risks, such as AI-enabled cybercrime, attacks by non-state actors, crisis communication, model safety testing and protection of critical infrastructure," Teng said. "Cooperation is not a concession; it is responsible self-protection."

liazhu@chinadailyusa.com