As China and the United States race to expand their artificial intelligence capabilities, economists said the two countries need closer cooperation on global governance and safety standards to better manage the technology's potential risks.

Although the US continues to lead in frontier AI research, large-scale models and computing power, economists said that China is rapidly catching up and has developed strengths in integrating AI into manufacturing, consumer services and other parts of the real economy, reflecting the two countries' different development priorities.

"There are two main players in the development of frontier AI — China and the US," economist and Nobel laureate Michael Spence told China Daily. "There's almost no measurable difference between them now in terms of performance. Whatever difference there was, China has caught up."

According to Stanford University's Institute for Human-Centered AI, since early 2025, US and Chinese AI models have traded places at the top of performance rankings multiple times.

According to the institute's 2026 AI Index Report, as of March, the top US model led by only 2.7 percent, with a marginal gap that has fluctuated in single digits over the past year.

Zhu Min, former deputy managing director of the International Monetary Fund, said, "The AI innovation models pursued by China and the US are actually quite different."

"The US focuses on large models, massive computing power and a strong foundational base, aiming for what we call 'artificial superintelligence'. What the US pursues is the development of intelligence itself," he said.

China, according to Zhu, measures AI development by a different yardstick.

"China, in contrast, emphasizes applying AI to the real economy. When assessing whether AI is good or not, China places greater weight on the benefits and convenience it can bring to the real economy and to people's daily lives," he added.

Edge in systematic approach

Spence noted that China's systematic approach to deployment gives it an edge that the West currently lacks.

"China has, in my view, in the context of the 15th Five-Year Plan (2026-30), a better set of either plans or intentions to make sure that AI is deployed by and adopted across a wide range of the economy, both in manufacturing and services," Spence said. "We talk a lot about the diffusion challenge, but we are just talking about it. We are not doing anything," he added.

As China, the US and other economies scramble to scale up their AI capabilities, the risks are increasing — not just economic ones.

Zhu warned, "AI carries significant uncertainty, including risks related to society, the environment and potential military uses."

"Therefore, establishing a governance framework and safety 'guardrails' for AI is, I believe, the most critical imperative for China and the US as the two leading players in this field, and it is also what the world expects of them," he added.

"The US tends to favor exogenous governance — building external oversight frameworks for monitoring and compliance," Zhu said. "China emphasizes an endogenous approach, embedding safeguards from the very beginning, even before model deployment, to ensure that AI benefits humanity."

Zhu noted that both approaches are valid and open to discussion, and could together shape the future of AI governance.

Earlier this month, Foreign Ministry spokesman Guo Jiakun announced at a news conference that China and the US had agreed to launch an intergovernmental dialogue on AI. Guo said that the two countries need to collaborate to promote AI development and improve its governance.

wangkeju@chinadaily.com.cn