Artificial intelligence (AI) should be developed with a people-centered approach and for the positive and good. This is China's principle on developing AI, outlined in a new white paper on China's international AI application and governance cooperation released on October 31.

The document calls for adhering to the principle of openness and win-win, ensuring that AI is secure and controllable, encouraging inclusiveness, and advocating solidarity.

It also introduces a shared technological foundation integrating models, computing power and corpus data. It traces how China's international AI application has evolved from single-technology and hardware exports to end-to-end industry solutions, and further to the inclusive sharing of open-source models, application tools, talent cultivation and governance experience.

The white paper then turns to practical cooperation, presenting typical cases between China and other countries, especially those in the Global South. Among them is a China-Brazil Science and Technology Backyard on family agricultural mechanization in Brasilia. The project applies the Internet of Things, AI and digital management to family farms, cutting agricultural materials and labor costs by 8 to 12 percent and raising per-unit yields of corn and sugarcane by 6 to 10 percent. In Apodi in northeastern Brazil, production efficiency climbed by 80 percent and farmers' incomes grew by 60 to 80 percent.

The health sector offers another example. Beijing Airdoc Technology, together with the Gates Foundation, a global health research institute and local medical institutions, developed a non-invasive AI fundus screening solution for maternal health needs in Africa, with low-cost camera modification, algorithms adapted to African facial features and weak-network environments, and offline use. The project has completed pilot implementation in Niger, with preliminary verification underway in Ghana, South Africa and Tanzania.

Meteorological disaster prevention is a further area of cooperation. The China Meteorological Administration's MAZU system is the world's first national-level AI meteorological public product responding to the UN's "Early Warnings for All" initiative. Supporting localized adaptation and lightweight cloud access, it has been localized in seven countries and is used by more than 40 countries via cloud access.

Talent cultivation has also seen concrete progress. The Beijing Academy of Artificial Intelligence, together with the African Union's African Scientific, Research and Innovation Council and Peking University, implemented a China-Africa AI computing talent capacity-building program. It covers university teachers from 10 African countries, with a long-term plan to reach 52. The program was recognized as an outstanding case of the 2026 AI for Good initiative.

In security and governance cooperation, Renmin University of China released the Foreign-Related Rule of Law Large Model 2.0, integrating legal rules, judicial precedents and governance practice data from major legal systems. It is open to members of the Global Law Education Cooperation Alliance. The university is working with nearly 30 leading law schools, including Oxford University, to conduct joint research on AI rule-of-law governance.

The white paper also proposes cooperation paths including sharing technological capabilities, livelihood-sector cooperation, security governance coordination and capacity-building.

Released at the main forum of the 2026 AI+ Conference in Beijing, it was compiled by the Gaoling School of Artificial Intelligence at Renmin University of China and the China-SCO Countries AI Application Cooperation Center. Twenty-four institutions, including the International Cooperation Center of the National Development and Reform Commission, participated as co-compilers.