On November 3, DSET met with the International Republican Institute (IRI) in Washington, D.C., for a strategic discussion on countering China’s AI global expansion through practical policy responses and international collaboration. The meeting brought together Adam King, IRI’s Regional Director for East Asia and the Pacific, and Caitlin Scott, Director of Countering Foreign Authoritarian Influence, with DSET’s Democratic Governance Program to explore implementation challenges and opportunities for leveraging IRI’s global democracy promotion network.
Kai-Shen Huang, Director of DSET’s Democratic Governance Program, presented findings on China’s application-driven AI development model, emphasizing that China’s strategy focuses on scaling AI products through everyday domestic use before export. Using the TikTok case, Huang outlined critical obstacles for democratic nations attempting to review Chinese AI imports: data review challenges stemming from China’s restrictive data laws, and algorithm verification difficulties arising from export controls. Huang noted that Chinese AI services are rapidly spreading globally through free or subsidized models, creating lock-in effects that make it prohibitively expensive for democracies to reclaim market share.
You-Hao Lai, Deputy Director of DSET’s Democratic Governance Program, presented key findings from his forthcoming report analyzing the data practices of ten major Chinese AI services. Lai’s analysis confirmed that many of these services store user data in China or facilitate access through intra-group data sharing and “lawful” government requests, ultimately granting the Chinese government broad access to user data. In response to these risks, Lai proposed three core policy recommendations for democratic governments:
- Ban the use of Chinese AI services in government agencies and critical infrastructure;
- Establish a formal import review mechanism for Chinese AI providers; and
- Develop coordinated cross-border data governance frameworks among democratic allies.
The discussion focused on practical implementation challenges and the strategic value of demonstrating proof of concept. King emphasized that China’s competitive advantage lies in prevalence over perfection. While Chinese AI may not be the strongest technically, its free availability through state subsidies enables rapid global spread and market lock-in, particularly across Africa and Latin America. Scott noted the current challenges in building consensus and momentum for AI regulation in democratic nations, observing that advancing effective AI governance requires addressing broader data privacy frameworks. Both King and Scott highlighted that successful policy implementation in Taiwan could serve as a powerful global example for other democracies, similar to how Australia’s approach to countering Chinese political influence became a widely referenced model.
