On the afternoon of September 28, Xue Lan, Distinguished Professor of Arts, Humanities and Social Sciences at Tsinghua University and Dean of Schwarzman College, delivered a keynote address at the 2026 Nishan Forum on World Civilizations to share his views on global AI governance.

In Xue's view, advancing global AI governance from conceptual consensus to institutional practice requires clear priority-setting and a clear distinction between matters of greater or lesser importance and urgency. He proposed six pathways for action. First, build the World Artificial Intelligence Cooperation Organization (WAICO) into a substantive and effective emerging multilateral platform, and leverage "mutual recognition of safety evaluations" and "open-source model capacity building" as entry points to foster public credibility. Second, leverage the key coordinating role of China and the United States as major countries, and adopt the "notification mechanism for risks associated with frontier AI models" as a priority theme for the first-round dialogue to forge fundamental strategic trust. Third, develop a tiered and coordinated governance architecture to resolve discrepancies surrounding "interface rules" across different tiers. Fourth, narrow the digital and AI divide through inclusive cooperation, unlocking "tangible public benefits for all," as exemplified by the MAZU (Multi-hazard, Alert, Zero-gap, Universal) intelligent meteorological early warning solution, to be rolled out across more countries. Fifth, promote diverse and inclusive development through mutual learning among civilizations, and harness open-source models to help countries build AI-enabled applications that align with local cultural and ethical norms. Sixth, reinforce the bottom line of security governance, advance technical collaboration such as mutual recognition of safety evaluations and alignment of standards, and preserve cooperation channels free from geopolitical interference for the scientific communities of all countries.