Managing Public Health with AI in Taiwan: The Roles of National Identity and Privacy Concern in Shaping Public Attitudes (Preprint)

<sec> <title>BACKGROUND</title> The COVID-19 pandemic accelerated the deployment of AI-driven surveillance systems for public health governance, raising both public health benefits and ethical concerns. In Taiwan’s post-pandemic context, public acceptance of government-deployed AI surveillance systems remains underexplored. </sec> <sec> <title>OBJECTIVE</title> This study investigates how perceived AI capabilities, privacy concerns, and national identity collectively influence public acceptance of AI surveillance technologies in Taiwan. </sec> <sec> <title>METHODS</title> We conducted a cross-sectional online survey in August 2024 with 4,899 adults in Taiwan, recruited through both university-affiliated and commercial platforms. The outcome variable was public acceptance of government-deployed surveillance AI systems. Key variables included perceived AI capability (independent variable), privacy concerns (mediator), and national identity (moderator), along with sociodemographic and political covariates. We applied Hayes’ PROCESS macro (Model 15) to test a moderated mediation model. All predictors were mean-centered, and 5,000 bootstrap samples were used to calculate 95% bias-corrected confidence intervals. </sec> <sec> <title>RESULTS</title> Higher perceived AI capability was associated with greater acceptance of government-deployed surveillance systems, both directly and indirectly through reduced privacy concerns. Taiwanese identity moderated this mediation pathway: individuals identifying strongly as Taiwanese exhibited higher acceptance, even when privacy concerns were high. These effects were more pronounced for intrusive AI surveillance technologies (e.g., geolocation, facial recognition) than for non-intrusive ones (e.g., health data analytics). </sec> <sec> <title>CONCLUSIONS</title> Public support for AI surveillance depends not only on perceived technical capability but also on sociopolitical context and collective identity. To ensure ethical AI governance, trust-building must be grounded in transparency, privacy protection, and public deliberation. </sec>