A GPT-based Approach for Research Article Identification: a Case Study in Artificial Intelligence

This study presents a comprehensive approach that addresses the challenges of identification and analysis of research articles in rapidly evolving fields, using the field of Artificial Intelligence (AI) as a case study. By combining search terms related to AI with the advanced language processing capabilities of generative pre-trained transformers (GPT), we developed a highly accurate method for identifying and analyzing AI-related articles in the Web of Science (WoS) database. Our multi-step approach included filtering articles based on WoS citation topics and category, keyword screening, and GPT classification. We evaluated the effectiveness of our method through precision and recall calculations, finding that our combined approach captured around 94% of AI-related articles in the entire WoS corpus with a precision of 90%. Following this, we analyzed the publication volume trends, revealing an increasing degree of interdisciplinarity. We conducted citation analysis on the top countries and institutions and identified common research themes using keyword analysis and GPT. This study demonstrates the potential of our approach as a tool for the accurate identification of scholarly articles, which is also capable of providing insights into the growth, interdisciplinary nature, and key players in a research area.