Key points are not available for this paper at this time.
In the context of digital government transformation, e-government AI assistants represent a critical component in the intelligent upgrading of government services, where understanding user adoption behavior becomes essential for enhancing service effectiveness. Building upon the extended UTAUT2 model, this study incorporates Trust in AI (TIAI) and Perceived Privacy Protection (PPP) to construct a comprehensive framework examining the adoption mechanisms of government service applications. The research specifically investigates how performance expectancy, effort expectancy, and social influence shape users' behavioral intentions and actual usage patterns. Through PLS-SEM analysis of 513 valid questionnaire responses, the study reveals several key findings. Social influence emerges as the most significant predictor of usage intention (β=0.277, p<0.001), while perceived privacy protection demonstrates particularly strong predictive power (β=0.235, p<0.001) along with the highest performance score (76.026). The analysis also identifies meaningful positive effects from effort expectancy (β=0.125, p<0.05) and habit (β=0.100, p<0.05) on usage intention. Regarding actual usage behavior, the results show significant influences from facilitating conditions (β=0.229, p<0.001), trust in AI (β=0.273, p<0.001), and usage intention itself (β=0.397, p<0.001). These findings not only contribute to the theoretical understanding of technology adoption within digital government transformation but also offer practical insights for optimizing the design and implementation of e-government AI assistants, revealing the dominance of social influence in collectivist contexts like China, privacy protection measures, and trust-building features in driving user adoption.
Li et al. (Fri,) studied this question.