Imagining the algorithmic city: Bias, cultural authenticity, and spatial futures in AI-generated representations of Indonesian urban environments

Abstract

Generative artificial intelligence increasingly mediates how Indonesian cities are visualized, yet its synthetic urban imagery may privilege iconic infrastructure and global metropolitan aesthetics while weakening social, linguistic, and cultural complexity. This study examines how algorithmic bias, cultural authenticity, and spatial futurity are constructed across AI-generated representations of Jakarta, Surabaya–Madura, Bandung, Yogyakarta, and national urban futures. A comparative qualitative multimodal design analyzes twelve static AI-generated images, supported by secondary scholarship and official contextual documents, through spatial representation auditing, a cultural-authenticity and semiotic-specificity matrix, and spatial-futures narrative analysis. Findings show that skylines, monuments, bridges, and engineered environments provide strong geographical recognizability, whereas informal spaces and socially differentiated actors remain comparatively marginal. Cultural identity is retained primarily through landmarks and religious architecture, but vernacular forms, linguistic signage, and everyday practices receive substantially less visibility. Future orientation is expressed mainly through technological order, polished infrastructure, and globally legible urban form, while ecological complexity and social inclusion remain limited. This study contributes an integrated analytical framework demonstrating that representational bias may operate through selective preservation rather than complete cultural erasure, thereby connecting urban visual analysis, cultural semiotics, and critical AI urbanism within a comparative Indonesian context and offering a transferable approach for evaluating synthetic cities.