Urban Public Art and Its Interaction Design in the Context of Artificial Intelligence

Generative AI is transforming urban public art in both creation and experience. This transformation extends beyond visual aesthetics because it changes the interactive logic among artworks, audiences, and urban space. This study focuses on AI-driven urban public art and its interaction design, examining how generative systems, affective computing, and sensor-mediated interaction reshape public artistic experience. Field cases and empirical data show that affective computing feedback mechanisms significantly drive audience engagement depth, with emotion-recognition accuracy showing a strong effect (β = 0.524, p < 0.001). AI intervention also influences urban spatial cognition, community cultural identity, and technology accessibility across multiple dimensions. Participants in the complete triadic interaction group demonstrate substantially deeper engagement than the control group (Cohen’s d = 1.24, 95% CI [1.02, 1.46]), and immersion increases by 60.9%. To mitigate demand characteristics, participants were unaware of the comparative hypotheses, and group assignment was randomized. The study further identifies systematic digital-divide disparities across age, educational attainment, and income. Based on these findings, a theoretical framework of AI-mediated publicness is proposed, providing empirical grounding for interaction design practice and policy governance of urban public art in the age of artificial intelligence.