Digital Innovation in Public Administration through Intelligent Public Sector Automation (IPSA): Strategies and Challenges

This study examines the implementation of Intelligent Public Sector Automation (IPSA) based on large language models (LLMs) for digital innovation in Korean public administration, focusing on critical challenges and their solutions. Through comprehensive literature review and case studies, the research identifies three fundamental challenges: the architectural differences between SQL-based systems and LLM systems in handling administrative data, security and access control requirements specific to LLM operations, and compliance with Korea’s Data 3 Laws (Personal Information Protection Act, Information and Communications Network Act, and Credit Information Act). The study analyzes both domestic and international cases, demonstrating that successful IPSA implementation requires robust data governance frameworks, sophisticated hybrid architectures, and comprehensive compliance mechanisms. Our research contributes by developing a novel framework for integrating LLM capabilities with traditional administrative systems while maintaining regulatory compliance, providing detailed technical specifications for secure and accurate IPSA systems, and establishing concrete guidelines for workplace transformation. The findings reveal that implementing IPSA requires careful consideration of data accuracy, security protocols, and privacy protection measures, particularly in the context of Korean regulatory requirements. This research provides valuable insights for public institutions adopting large-scale AI technologies, offering specific guidelines for implementing IPSA systems that meet both operational requirements and regulatory obligations, while suggesting directions for future empirical research on IPSA effectiveness and implementation strategies in Korean public administration.