Arts in transit under generative AI: Visual governance and the reconfiguration of publicness in Beijing’s metro space

Abstract

As arts in transit has become an increasingly prominent practice in urban public transportation spaces worldwide, it is important to understand how AI-generated images are reshaping visual orders across different urban contexts. Taking Beijing Metro Line 2 as the primary case, supplemented by contextual observations on Lines 1, 4, and 10, this study draws on participant observation and semi-structured interviews. The findings reveal three interconnected changes. First, traditional transit art continues to provide durable cultural references, yet it faces growing competition from rapidly renewed visual content for limited passenger attention, producing a form of attentional displacement. Second, generative AI accelerates an already commercialized visual regime by lowering production costs, shortening revision cycles, increasing scalability, and redistributing authorship across algorithms, designers, and managers. Third, AI-generated imagery has expanded from advertising into public-service communication and cultural content, where it is incorporated into existing processes of institutional review and display. These changes reveal two dimensions of visual governance: institutional control over the production and display of visual content, and the organization of attention and spatial meaning through visual means. The Beijing Metro case therefore reconsiders the public value of arts in transit by showing how AI-driven visual production is reconfiguring the institutional, cultural, and perceptual conditions through which publicness is produced in urban transit space.

The authors' abstract, as published at the source. Geoforum, 2026 · DOI ↗