As the global landscape of Artificial Intelligence (AI) governance rapidly evolves, research has increasingly moved beyond state-centric perspectives to investigate the role of non-state actors. This paper focuses on an underexplored category of such actors: think tanks, specifically two prominent Chinese institutions—the China Academy of Information and Communications Technology (CAICT) and the Centre for International Security and Strategy (CISS). Drawing on the Communities of Practice (CoPs) approach, this study investigates how these think tanks shape the emerging governance framework for AI technologies by leveraging their position at the boundaries between various CoPs. Specifically, this position of in-betweenness enables them to influence AI governance through acting as boundary brokers. They bridge different Chinese CoPs—including the government, the private sector, and academia—and, in some cases, international CoPs, facilitating engagement and exchange across these communities. This paper also finds that due to differences in their organisational types and relationships with the Chinese government, CAICT and CISS influence AI governance in subtly distinct ways. This paper contributes to CoP scholarship by examining its applicability in an authoritarian context and is among the first to provide a timely empirical analysis of the role of Chinese think tanks in AI governance.
Photo by Alison M