The evolution and current landscape of AI in geographical research: A large-scale systematic review

Abstract

With the rapid advancement of Artificial Intelligence (AI) technologies, its applications have become increasingly widespread across various aspects of geography, offering unprecedented analytical capabilities across disciplinary boundaries. Despite this revolutionary potential, a comprehensive understanding of the current research landscape and development trajectory of AI in geographical sciences remains limited. To fill this gap, we conducted a large-scale systematic review based on 400,000 geographical publications published from 1990 to 2023. We utilized large language model (LLM) prompt engineering, topic modeling and other natural language processing techniques to analyze the publications. Our findings reveal that AI applications constitute 8.1 % of geographical research, with publication volume having increased 20-fold over three decades. Both China and the United States have been the leading contributors to AI-driven geographical studies, together accounting for 62.78 % of all publications in this field. Notably, more than half of the studies used traditional machine learning methods. Among the various geographical topics, remote sensing applications and spatial data analysis emerged as the most extensively explored areas using AI techniques, with image feature extraction being the topic with the deepest level of adoption and most significant ongoing impact of AI methods. This systematic review provides critical insights into the integration trajectory of AI within geographical sciences, establishing a foundation for identifying emerging research opportunities and enhancing our understanding of AI’s transformative role in advancing geographical knowledge.