Geopolitics of AI Adoption

This project examines whether people are willing to adopt AI assistants developed by authoritarian states, and whether the competitive advantages of such assistants – lower cost and superior performance – can offset reluctance toward their country of origin. We conduct a large-scale cross-national conjoint experiment across ten countries: the United States, United Kingdom, Germany, Italy, Czech Republic, South Korea, Mexico, India, Nigeria, and South Africa. The sample is designed to capture variation in geopolitical orientation toward China, levels of AI adoption, regional diversity, and political regimes. In each conjoint task, respondents are presented with profiles of two hypothetical AI assistants that vary across nine attributes: country of origin (United States or China), cost, and performance in seven common use cases – text writing and editing, code writing and editing, answering user questions, searching for information online, and image, video, and audio generation. Respondents indicate which AI assistant they would prefer to use and separately rate each assistant on a 1–10 likelihood-of-use scale. We estimate average marginal component effects (AMCEs) to identify whether there is a preference penalty for AI assistants developed in China, and average component interaction effects (ACIEs) to examine whether superior performance or lower cost can offset this penalty. The study addresses two main questions: whether the authoritarian origin of an AI assistant shapes user preferences across different geopolitical contexts, and whether price and performance trade-offs are sufficient to overcome such concerns. The findings contribute to debates on digital authoritarianism and the diffusion of authoritarian-origin technology, and speak to policy discussions on the regulation of foreign AI assistants in democratic societies.