A Chinese policy scholar has called for a global artificial intelligence governance system that combines international rule-setting, broader access to AI capabilities and coordinated safeguards against emerging risks.
The proposal was presented in a commentary published during the 2026 World Artificial Intelligence Conference and High-Level Meeting on Global AI Governance, held in China from July 17 to 20. Ding Jie, director of the International Relations Center at the Academy of Contemporary China and World Studies, argued that AI should be governed through an inclusive multilateral framework rather than through competing national blocs or exclusive technology alliances.
The commentary described China’s preferred approach as one that promotes innovation while placing risk controls throughout research, deployment and international cooperation. Its stated principles include human oversight, early-warning systems, emergency-response mechanisms, transparency and shared access to the benefits of AI.
The proposal arrives as governments pursue different approaches to regulating increasingly capable AI systems. Some jurisdictions have adopted or proposed legally binding rules based on risk categories, while others rely more heavily on voluntary commitments, technical standards and existing sector-specific laws. Those differences have complicated efforts to establish common requirements for safety testing, accountability and cross-border deployment.
Ding argued that global governance should address both the risks of AI and unequal access to the technology. The commentary highlighted China’s pledge to provide 5,000 specialized AI training opportunities to developing countries over five years, presenting the initiative as a way to narrow gaps in technical skills and institutional capacity.
It also pointed to China’s Global AI Governance Initiative and its role in AI-related resolutions adopted by the United Nations General Assembly. UN resolutions can establish shared political principles and encourage cooperation, although they generally do not create the kind of enforceable obligations found in domestic legislation or binding international treaties.
The commentary said a World Organisation for AI Governance was inaugurated in Shanghai during the conference with 29 founding members, describing it as an intergovernmental body intended to support multilateral cooperation. It did not provide detailed information about the organization’s charter, membership criteria, legal status or enforcement powers. Those elements would determine whether the body develops into a rule-setting institution, a coordination forum or a venue for voluntary initiatives.
Ding also cited the deployment of a Chinese-developed intelligent early-warning system called Mazu in 30 countries as an example of practical cooperation on risk prevention. The commentary did not specify the system’s applications, performance measures or oversight arrangements.
Open-source development was another central part of the proposal. Ding presented models released by Chinese companies during the conference as an alternative to concentrated control of advanced AI technology. Open models can expand access for researchers, businesses and governments, but they also raise governance questions because software and model weights may be adapted or redistributed beyond the control of their original developers.
The framework reflects a wider debate over whether AI governance should be led primarily by national governments, existing international institutions, new specialized bodies or technical organizations. Any global arrangement would also need to reconcile disagreements over privacy, security, intellectual property, content controls and the acceptable uses of AI.
The proposal places particular emphasis on participation by developing countries, which often have limited representation in technical standard-setting despite facing the effects of systems built elsewhere. Its next test will be whether the conference initiatives produce clearly defined institutions, transparent procedures and commitments that attract support beyond countries already aligned with China’s approach.
Sources: AI regulation