AI Adoption in Public Management and Leadership

Session Submission Type: Panel

Abstract

This panel explores the evolving role of leadership, bureaucratic capacity, and governance structures in shaping how AI is integrated into public management, impacting the landscape of administration and service delivery with an emphasis on effective digital governance and equitable AI adoption.  In particular, the authors examine the adoption of AI in public management as not merely a technological process, but rather a multiscale phenomenon framed by the interactions between institutional structures and especially individual actors.  The panel presents three empirical studies that collectively examine the sociotechnical systems through which AI becomes embedded in the operations of government: across national legislatures, municipal governments, and street-level bureaucracy.

At the national level, the emergence of comprehensive AI regulation in South Korea (culminating in the passage of the AI Act in 2024) provides a unique opportunity to study the institutional preferences and tensions that underlie formal governance structures. Using a conjoint experimental design with members of South Korea’s National Assembly and affiliated staff, the study examines six key dimensions of AI governance—regulatory formality, stakeholder inclusion, decentralization, public engagement, international cooperation, and agency responsibility. The results highlight competing logics: a centralizing impulse favoring regulatory efficiency and control versus a participatory impulse emphasizing transparency and legitimacy. These tensions echo broader global debates about balancing innovation, economic competitiveness, and ethical safeguards.

At the subnational level, a second study investigates the determinants of AI adoption across 100 U.S. city governments. The focus is on two variables, the political ideology of mayors and the professional background of Chief Information Officers (CIOs). Drawing on a newly compiled dataset of municipal AI initiatives between 2018 and 2024, the study finds systematic differences in adoption patterns as a function of leadership traits. Conservative versus liberal mayors, for example, display differential emphasis on public safety versus social service applications, while CIOs with technical training and public sector experience are more likely to pursue integrated, mission-aligned deployments. These findings suggest that city-level AI use is a function not only of resource constraints or technical capacity, but also of political ideology and the backgrounds of administrative professionals.

Complementing these organizational and institutional perspectives is a third study focusing on the individual-level dispositions of bureaucrats involved in public service delivery. Through two vignette experiments based on real-world scenarios (tax filing assistance and emergency rental programs), this study assesses how bureaucrats evaluate AI-enabled service delivery. Findings show that support for AI is influenced by public service motivation (PSM), perceived administrative efficacy, and the specific policy domain in which AI is deployed. Importantly, bureaucrats are shown not as passive implementers but as critical actors whose interpretations and values shape the integration of AI within existing workflows and public values.

Policy Area

• Public and Nonprofit Management and Finance

Secondary Policy Area

• Innovations in Science and Technology

Chair

• Gregory Dobler, University of Delaware

Discussant

• Christina Walker, Purdue University

Organizer

• Gregory Dobler, University of Delaware

Individual Presentations

• Governing AI: Legislative Preferences and Challenges in Korea’s National Assembly - Presenting Author: Chee Hae Chung, Purdue University; Non-Presenting Co-Author: Daniel S. Schiff, Purdue University

• Municipal Leadership and AI Adoption - Presenting Author: Heonuk Ha, University of Michigan; Non-Presenting Co-Author: Kaylyn Jackson Schiff, Purdue University; Non-Presenting Co-Author: Ogadinma Enwereazu, Purdue University; Non-Presenting Co-Author: Yunzhe Liu, Purdue University

• Bureaucratic Perspectives on AI in Public Service Delivery: An Experimental Approach - Non-Presenting Co-Author: Donavon Johnson, Florida International University; Presenting Author: Lerone E Laing, Florida International University

• Representation Effect in Nonprofit Government Grants: AI Webinars and Grant Applications - Presenting Author: Jongmin Lee, Indiana University