Research on the Construction of a Citizen-Participatory Design Support Methodology for the Semantic Interpretation of Urban Spaces Using Multimodal AI

This study proposes a dialog-based multimodal AI system to support citizen-participatory urban planning by integrating spatial information and citizen opinions. The system combines urban attribute data and street-level images with interactive dialogue to incrementally structure subjective and ambiguous citizen feedback into actionable planning issues and policy options. A user study conducted in the Sannomiya area of Kobe City demonstrated that the system can operate within practical time constraints and effectively clarify users’ concerns and underlying conditions. The results suggest that the proposed approach has potential to support early-stage consensus-building and policy exploration in participatory urban planning processes.