Can Digital Governance Training Harmonize Civil Servants’ Perceptions of AI Decision-Making? A Pre and Post-Test Design
Hsini Huang, Don‐Yun Chen, Yu-Han Chen · Information Polity · 2025
Successful AI adoption and implementation in public organizations require new technological readiness and recalibration of civil servants’ attitudes and perceptions toward AI. This study utilizes a pretest-posttest survey design to assess the impact of digital governance training on Taiwanese civil servants’ perceptions of AI discretion, specifically examining perceptions of legality, efficiency, and fairness. Findings of this research identify diverse AI attitudes among public employees—optimistic, pessimistic, doubtful, and pragmatic—and assess how these attitudes influence their views on AI's legitimacy. Digital governance training serves as a strategic tool to recalibrate civil servants’ perceptions, particularly effective in mitigating pessimistic views. However, it can also intensify existing beliefs. Optimists tend to view AI's efficiency more positively post-training, whereas doubters become more skeptical about its legality. Overall, the findings suggest that training variably impacts civil servants’ perceptions of AI, highlighting that while it is not a quick remedy, it is a critical tool in managing and shaping administrative attitudes toward AI.