Applications of artificial intelligence in public health: analyzing the built environment and addressing spatial inequities

Aim: To review the application of artificial intelligence (AI), specifically computer vision, in analyzing built environment (BE) characteristics within public health research, with a focus on spatial equity. Subject and methods: We conducted a rapid review of peer-reviewed articles (2014–2024) in English that integrated AI or computer vision in public health research on the BE. Following JBI and PRISMA guidelines, with a registered PROSPERO protocol, we searched Web of Science, PubMed, and Scopus databases. Data were extracted using a JBI-adapted template and synthesized descriptively, focusing on methods, key findings, and spatial equity elements. Results: Ten cross-sectional studies, predominantly from urban areas in the USA and China, met the inclusion criteria. These studies used computer vision to analyze BE features such as roads, greenery, and buildings through street view or satellite images. Health outcomes examined included physical activity, mental health, obesity, and mortality. Findings consistently showed positive health associations with increased greenery and improved street infrastructure. However, spatial equity was minimally addressed, with only one study (10%) considering this aspect. Conclusion: While AI applications in public health research on the BE show promise, there is a need for further research to address spatial equity and ensure findings are inclusive and relevant across diverse populations and contexts.