A Study on the AI Governmance Model for Policy Design Based on Voluntary Communities of Practice(CoP) : Multi-layered Structural Design of an Intelligent Policy Lab for Outdoor Advertising Administration

(Background and Purpose) As contemporary urban social issues evolve into “wicked problems” characterized by high complexity and interdepartmental silos, traditional bureaucratic structures face critical limitations in flexibility, continuity, and technical expertise. This study explores the fundamental necessity of introducing 'Policy Design' through voluntary cooperation networks within public institutions to overcome administrative fragmentation and the chronic disconnect of expertise caused by frequent personnel rotations. The primary purpose is to establish a future-oriented administrative framework that ensures consistency and public value in high-risk administrative areas such as outdoor advertising.(Methodology) This research employs a multi-faceted approach, beginning with a comprehensive literature review of Christian Bason's 'Public Design' theory and Ben Shneiderman's 'Human-Centered AI (HCAI)' framework. It further incorporates Wenger's Community of Practice (CoP) theory to analyze the structural dynamics of voluntary networks. Furthermore, it analyzes the operational data of a voluntary network for outdoor advertising officials, which has achieved a high satisfaction rate of 93.0%. The study specifically focuses on the structural design of an 'Intelligent Policy Lab' that integrates human expertise with AI-driven automation.(Results) The study proposes a multi-layered structural model for an Intelligent Policy Lab consisting of three levels: Human (Assetization of expertise), AI (System intelligence), and Network (Governance). A key finding is the efficacy of the “Data Feedback Loop,” where weekly query-response data from voluntary networks is archived and used to train AI models for automated responses. This process ensures administrative consistency and significantly reduces the cognitive burden on officials. Moreover, the application of the HCAI framework allows for high automation in repetitive tasks while maintaining high human control for complex value judgments.(Conclusions) In conclusion, the transition to “Design Administration” through intelligent policy labs is essential for enhancing the quality of urban environments. This research suggests that institutionalizing voluntary networks as official innovation bodies can bridge the gap between policy design and execution. Policy recommendations include the establishment of a national-level public design policy lab and the legislative integration of ethical governance in AI implementation. Future research should develop standardized performance metrics for AI-assisted administrative systems and conduct longitudinal studies to measure their impact on citizen-perceived public value and urban landscape quality.