Abstract

Generative artificial intelligence (GenAI) is widely considered to hold transformative potential for public administration, but its adoption in local governments remains uneven, weakly institutionalized, and shaped by informal employee experimentation. This study advances the literature by examining GenAI adoption during a rare moment of institutional restructuring in Vietnam, where the 2025 administrative reform abolished the district tier and shifted responsibilities to ward- and commune-level government. To address the unique dynamic of GenAI adoption, this research extended the Unified Theory of Acceptance and Use of Technology (UTAUT) by integrating trust in technology, perceived risk, technological awareness, and perceived organizational support. Survey data were collected from 302 civil servants across eight post-merger ward- and commune-level administrative units in a transitioning Vietnamese province. The study employs a dual-methodological approach, integrating Partial Least Squares Structural Equation Modeling (PLS-SEM) with fuzzy-set Qualitative Comparative Analysis (fsQCA) to evaluate both net effects of individual variables and the underlying causal complexity. The PLS-SEM results show that performance expectancy, social influence, technological awareness, and perceived risk are positively associated with GenAI adoption intention, while effort expectancy and trust in technology operate indirectly through performance expectancy rather than exerting a direct effect. This suggests that, in the context of local government reform, trust matters primarily when it strengthens employees’ belief that GenAI can improve work performance. Perceived organizational support attenuates the positive relationship between perceived risk and adoption intention, suggesting a buffering role in reducing risk-driven experimentation under institutional uncertainty toward more structured use. The fsQCA findings reveal causal asymmetry and equifinality: adoption intention emerges through multiple pathways anchored in perceived usefulness, social influence, and facilitating conditions, whereas non-adoption is associated with low awareness and weak organizational support. The study contributes to public administration, digital governance and technology acceptance research by showing how GenAI adoption is shaped not only by individual perceptions, but also by institutional change.

IPC Classification

Keywords