Abstract
Artificial intelligence geography, as a frontier interdisciplinary field integrating geography and artificial intelligence, is driving a systematic transformation in spatial cognition paradigms, analytical methodologies, and decision-making models. Based on a comprehensive review of current knowledge, this study constructs a theoretical framework of artificial intelligence geography, highlighting advances in core technologies such as intelligent processing of geospatial data, spatial prediction and simulation, and spatial optimization and intelligent decision-making, while summarizing representative applications in urban governance, environmental sensing, social spatial behavior analysis, and disaster management. Addressing key bottlenecks in current research—including insufficient capability in spatial causal inference, limited model interpretability and trustworthiness, challenges in capturing cross-scale consistency, and concerns over data privacy and algorithmic fairness—this study provides a critical analysis from the perspectives of integrating data-driven paradigms with geographic process mechanisms, and advancing geographic knowledge representation and reasoning. On this basis, future research directions are proposed, including developing foundation models with stronger geographic constraints, promoting interpretable and trustworthy geographic artificial intelligence technologies, constructing planetary-scale geographic intelligence systems, and fostering AI-empowered participatory geography and sustainable development decision-making. This work aims to provide a systematic reference and research implications for theoretical enhancement, technological innovation, and strategic applications of artificial intelligence geography.
