The Applications of AI on Urban Environment-Health Research: Insights Across Multiple Spatial Scales

Artificial Intelligence (AI) provides robust technical support for capturing, analyzing, and modeling the intricate relationship between the urban environment (UE) and residents’ health. This article systematically reviewes the related research published from 2017 to 2024 and presents two analytical frameworks. The first framework delineates AI-driven extraction methods for UE indicators across multiple spatial scales. The study reveals that AI is predominantly utilized at the microscale, particularly in the extraction of environmental variables from image data. The second framework divides the urban environment-health (UE-health) research into two distinct domains based on the research objectives: association exploration and mechanism analysis. This review shows that AI is widely applied in association exploration, facilitating variable construction or predictive modeling. In contrast, mechanism analysis still predominantly relies on traditional methods. Despite substantial gains in efficiency and data coverage, AI applications in UE-health research remain constrained by heterogeneous data sources that limit comparability across studies, semantic inconsistencies across scales that impede transferable modeling, narrow coverage of health outcomes, and limited model interpretability that restricts mechanism-oriented inference. Addressing these challenges will require the development of transferable, semantically consistent indicator systems, the integration of AI with causal inference and theoretical frameworks, and the implementation of dynamic, data-driven, scale-aware models. Advancing along these lines could shift the field from predominantly association-driven studies toward mechanistic understanding and theory-informed applications of AI in UE-health research.