On May 5, 2026, the Democratic Governance Program at the Research Institute for Democracy, Society and Emerging Technology (DSET) visited the Jamestown Foundation for an exchange with President Peter Mattis, Senior Fellow Matthew Johnson, and other researchers from the Foundation. The discussion focused on the global expansion of Chinese AI capabilities, emerging technology competition, and foreign information manipulation and interference.
Kai-Shen Huang, Director of DSET’s Democratic Governance Program, presented his recent research analyzing how the capabilities of China’s AI ecosystem are advancing through several key vectors, including open-weight models, open architectures, and modular components. He underscored the structural significance of this trend: the diffusion of Chinese AI capabilities is not the product of a single state strategy, but rather the adaptive evolution of China’s AI ecosystem under sustained export control pressure, which existing chip controls and policy tools are insufficient to address. Violet (Yueh-Ning) Chiang, Non-Resident Fellow at DSET, introduced the team’s ongoing project on FIMI in AI, including its methodology and an in-house trained censorship-detection classifier, and explained how the tool is being used to examine differences in major large language models’ responses to politically sensitive topics across different language environments. You-Hao Lai, Deputy Director of the Democratic Governance Program, gave a brief overview of the data governance structure analysis developed in The Authoritarian Gaze report and engaged in in-depth discussion with Senior Fellow Matthew Johnson, whose research has long focused on China’s cross-border data acquisition.
The Jamestown Foundation shared its current research priorities, including AI robotics, quantum technologies, energy security, brain-computer interfaces, and the connected-devices supply chain research led by Matthew Johnson. Several of these areas overlap substantively with DSET’s ongoing work. Mattis emphasized that competition with China in emerging technologies must be understood within the broader frame of political competition, and that technical analysis must remain tightly connected to the strategic intent of the Chinese party-state in order to inform effective policy discourse. The two sides further discussed the competitive and cooperative dynamics among Chinese AI firms, the implications of open-source model proliferation for the global information environment, and specific corporate cases such as GoLaxy that illustrate the scope of foreign information manipulation capabilities.
