<p>Geological safety (geosafety) refers to the state in which human settlements and engineering activities are protected from threats posed by geological processes. With the intensification of climate change and human engineering activities, geosafety risks have emerged as critical challenges to the construction of mega projects and the safe operation of urban areas in China. The rapid accumulation of massive, long-term, multi-source spatiotemporal geoscience data offers essential insights into historical and potential geosafety issues, providing a solid foundation for geosafety research. As an emerging tool for data-driven analysis, artificial intelligence (AI) has significantly advanced the understanding and management of geosafety issues in both mega projects and urban areas. This paper reviews recent advances in applying AI to common geosafety issues, with a particular emphasis on applications in China’s mega projects and urban areas. First, we outline the motivation for AI-based geosafety research, including the growing limitations of traditional approaches and the potential of AI-based solutions. We then introduce an overview of applying AI approaches for geosafety analysis in mega projects and urban areas. This paper further investigates the application of AI in addressing major geosafety issues in mega projects, such as the Sichuan-Tibet Railway, focusing on earthquake-induced landslides, thaw slumps, glacier debris flows, and deep tunnel risks, including rockbursts and large deformations. In the Three Gorges Reservoir area, rainfall-induced landslides are emphasized. For urban environments, attention is given to ground subsidence, ground collapse, and the recent surge of clustered landslides triggered by extreme rainfall. Finally, several potential challenges and future directions of AI for geosafety research are discussed.</p>