한국지방행정연구원

This study aims to comprehensively examine the administrative environment and policy conditions faced by local governments in the era of artificial intelligence (AI) transition and to propose innovation strategies for local administration based on resident-centered AI governance that enable the effective implementation of national public-sector AI policies at the local level. The rapid diffusion of generative AI and the acceleration of digital transformation are fundamentally reshaping the provision of public services and policy decision-making processes, elevating the use of AI in the public sector from a technical option to a core government policy agenda. At the same time, growing societal demands for accountability, transparency, and trust have accompanied the expanded use of data and algorithms in public administration. However, significant disparities among local governments in terms of AI adoption levels, institutional arrangements, organizational capacity, and fiscal resources have constrained the uniform implementation of central government policies at the local level.
To address these challenges, this study reviews domestic and international theories and policy discussions on AI-driven administrative innovation and public-sector AI governance, establishing a conceptual framework of resident-centered AI governance grounded in explainability, accountability, transparency, and participation. It then analyzes the legal and institutional environment and the current status of AI adoption across local governments in Korea, drawing on nationwide surveys. The analysis reveals that although legal frameworks such as the Artificial Intelligence Basic Act are being developed, operational standards related to accountability, risk assessment, and explainability remain insufficiently specified, resulting in substantial variation in policy implementation across jurisdictions. Disparities are also evident in the establishment of dedicated organizations, local ordinances, data infrastructure, and human and financial resources, with many basic local governments remaining at an experimental or partial stage of AI utilization.
In addition, survey results of citizens and public officials indicate that expectations regarding administrative efficiency and service quality coexist with strong concerns over data privacy, algorithmic bias, and unclear responsibility in cases of malfunction or error. Differences in trust and policy acceptance are closely associated with age, digital competence, and experience with generative AI, while public officials tend to adopt a cautious stance toward AI implementation due to limited expertise and increased accountability burdens. These findings underscore that AI adoption in local government constitutes a complex policy challenge that extends beyond technical feasibility to encompass social trust, organizational capacity, and institutional readiness.
Based on these findings, the study proposes a comprehensive direction for AI-driven transformation in local administration, emphasizing a shift from fragmented and reactive practices toward data-driven, connected, and anticipatory governance. It highlights the role of local governments not merely as adopters of technology but as active policy experimenters and coordinators. By articulating institutional and operational principles for resident-centered AI governance and suggesting phased and typology- based implementation roadmaps, this study provides analytical and policy foundations for aligning national AI policy design with local-level execution. The findings may serve as a basis for developing standard local ordinances, refining tailored policy models for different types of local governments, and supporting future research on sustainable AI-enabled administrative innovation.