Explainable artificial intelligence for urban well-being in sustainable city ecosystems

Abstract Urban well-being has been established as a central performance indicator of sustainable city ecosystems, while conventional AI models are considered insufficient in terms of transparency, accountability, and interpretability. The purpose of this study was to synthesise empirical evidence on XAI for urban well-being within sustainable city ecosystems. Included studies were identified using the PRISMA 2020 flow framework, with data retrieved from Scopus, Web of Science, IEEE Xplore, and the ACM Digital Library between 1 January 2020 and 1 November 2025. From an initial corpus of 49,372 records, 8914 were excluded, 3572 underwent full-text assessment, and 726 were included in the meta-analysis. Our benchmarking meta-analysis showed that environmental health, infrastructural resilience, mobility, safety, and social well-being have a positive association with XAI in enhancing urban well-being within sustainable city ecosystems. The implications were underscored, with transparency reaffirmed as a necessary condition for the sustainable deployment of XAI in support of urban well-being and resilient systems within liveable cities across Africa, South Asia, Southeast Asia, and Latin America.