China Promotes Security-Driven AI Governance
China is actively promoting a security-driven AI governance model within ASEAN and the Global South, based on its Global Security Initiative. The approach prioritizes "AI security," social stability, and data localization over Western models focused on individual rights. This push is supported by a planned vast AI infrastructure buildout in Xinjiang, aimed at bypassing export controls.
- China's Global AI Governance Initiative, proposed by Xi Jinping in October 2023, advocates for a "people-centered and AI for good" approach, emphasizing the increased representation of developing countries in the global AI governance framework. This initiative is part of a larger set of frameworks including the Global Development Initiative (GDI), Global Security Initiative (GSI), and Global Civilization Initiative (GCI). - The promotion of this governance model in the Global South is positioned as an alternative to Western, rights-based approaches, offering a path founded on "material pragmatism" and tangible utility in sectors like agriculture and healthcare. This resonates with the developmental priorities and state-centric governance ethos of many Southeast Asian nations. - In December 2025, China proposed a broad initiative to deepen digital governance cooperation with ASEAN, covering cybersecurity, cross-border data flows, and joint development of AI security standards and policies. This includes sharing an AI-powered phishing detection system and establishing a China-ASEAN cybersecurity threat intelligence platform. - The EU's AI Act, which takes a risk-based approach to regulation, and the US's market-driven model with its AI Bill of Rights, present contrasting frameworks to China's state-led model. While the EU and US focus on individual rights and safety, China's regulations, such as the 2021 rules on recommendation algorithms, prioritize social stability and adherence to socialist values. - The Xinjiang data center buildout is substantial, with plans for 39 facilities expecting to use over 115,000 Nvidia H100 or H200 chips, which are subject to U.S. export controls. A significant portion of this capacity, 70%, is planned for a single government-backed complex near Yiwu, which will benefit from incentives like free electricity. - Despite U.S. export controls, Chinese firms are leveraging older deep-ultraviolet immersion (DUVi) lithography tools with advanced multipatterning techniques to produce AI chips that are nearing cutting-edge capabilities. In 2024, Chinese buyers purchased 70% of the dominant DUVi system manufacturer ASML's machines. - China's domestic AI regulations are not purely top-down; major tech companies like ByteDance and DeepSeek contribute to shaping the guardrails through self-regulation and commercial strategies. This reflects an interaction between state authority, market incentives, and societal actors. - In July 2025, Chinese Premier Li Qiang proposed the establishment of a World Artificial Intelligence Cooperation Organization (WAICO) to coordinate global standards and ethical principles, positioning China as a key player in international AI norm-setting. This aligns with the "Top 10 Technology Buzzwords of 2025" list from the Chinese Society for Science and Technology Journalism, which was topped by "AI governance."
