Reading Urban Open Spaces: How Does Artificial Intelligence Compare Human Experts?

This study examines the extent to which generative artificial intelligence aligns with human experts in the visual assessment of urban open spaces. Site quality is addressed across five dimensions: spatial accessibility and organization; usage and social function; environmental quality and comfort; safety and inclusivity; and perceptual and aesthetic quality. The same images, indicators, and five-point scale regarding three heavily used public open spaces in Rize city center were evaluated by 20 architects and landscape architects, as well as by ChatGPT and Gemini. The comparison revealed that the level of agreement between the AI and the experts varied depending on the specific site and quality dimension. While alignment was high regarding features that are directly observable—both physically and visually—differences became more pronounced regarding perceptual and aesthetic qualities. The findings suggest that AI can serve as a complementary tool to support visual assessment rather than a replacement for human expertise.