國立臺灣大學政治科學論叢

The growth and breakthroughs in innovative artificial intelligence (AI) technologies have provided many governments and public organizations with the possibility of achieving digital transformation. Yet, apart from the proposals of techno-optimists, as more problems emerge in implementing AI in society, politics, and business, there are calls for scholarly discussions to revisit how AI might reshape and influence government values, bureaucracy, and bureaucratic operations. This paper draws from both the Weberian ideal bureaucracy theory and the organizational humanism perspective as an upper-level theoretical framework to explore the impact of AI on the roles, control, and communication issues in public organizations. As a stepping stone towards AI research in government, this study aims to analyze public administration theory and the applications of AI in public sector practices. Three key issues are discussed from a multi-level analytical framework, including: 1) the problem of public value misalignment, 2) the problem of multiple agents within an organization, and 3) the issue of AI discretion and automation bias. This article concludes that when facing the inevitable AI revolution, public organizations should rethink the foundations of innovative change, path selection, talent management, and the “re-design” of divisions of labor.