China’s AI governance action plan

Unlike the U.S. Action Plan, which focuses on preserving U.S. national interests, the Chinese plan strongly emphasizes global cooperation as the best path forward.

Photo by Saradasish Pradhan on Unsplash

China released its Global AI Governance Action Plan on July 26, 2025, just three days after the U.S. published its own AI Action Plan. The release coincided with the kickoff of China’s World AI Conference – a government-sponsored annual showcase of (mostly Chinese) AI companies which also attracts researchers, investors and government officials from around the world.  

Themes of China’s AI action plan

Unlike the U.S. Action Plan, which focuses on preserving U.S. national interests, the Chinese AI action plan strongly emphasizes global cooperation as the best path forward. It calls for the United Nations (UN) to play a leading role in monitoring AI developments and facilitating inclusive global discussions and explicitly supports diverse UN plans and bodies, such as the Global Digital Compact and the International Telecommunications Union (ITU). 

Additional themes include: 

• Sustainability | The Action Plan advocates for “global clean power,” “green computing technologies" and more global discussion of how AI can be used to address its own environmental impacts. 

• Standardization and safety | The Chinese government calls for standards for AI energy and water efficiency, computing power, technical specifications, data privacy, data security, and R&D. International bodies like the International Organization for Standardization (ISO) should facilitate discussion while industry formulates appropriate standards. Doing so can help improve safety assessment and threat response.

• Openness | Openness is also seen as necessary for threat anticipation and response as well as for rapidly advancing AI research and application. The Action Plan promotes openness in research collaboration (e.g., information sharing  platforms) and in dialogue (e.g., best practice dialogues among inter/national government officials). It also promotes open source AI and technological and governance interoperability.

• Multistakeholderism | AI governance dialogues should include all relevant stakeholders. As in other areas of technology policy, specific roles are presented for all types of actors, including governments (public sector integration, AI governance), industry (application and standard development), scientists (innovation, information sharing, risk assessment), and international organizations (facilitate dialogue, engage and assist the Global South). 

• Efficient innovation | Efficiency and innovation are necessary to contain the environmental and technological risks of AI and to achieve other international policy goals, such as facilitating economic development in the Global South, constructing digital infrastructure worldwide, and addressing climate change.

Cascade's take

In China's AI action plan, the Chinese government is pulling out all the stops to develop a structural advantage in the AI race. It aims to solidify its technology in global standards and build as many allies as possible through a mix of openness and generous development policy. Its focus on energy efficiency means China stands to minimize the impact of AI on the electric grid.

The Chinese government is applying the lessons it learned from the global race to 3G and then 5G global network dominance. If China was caught flat-footed by the build-out of 3G, it did not make the same mistake with 5G, where it is arguably winning the global race for dominance.

The U.S. is leading right now in innovation, but China is making strong strides in AI. If the U.S. steps back from global engagement, China is ready to fill that gap.