The development of Artificial Intelligence (AI) in higher education has grown significantly over the past decade, particularly following the emergence of generative AI technologies such as ChatGPT. This study aims to map the evolution of AI research in higher education from the pre–generative AI era to the emergence of ChatGPT using a bibliometric approach. Data were collected from the Scopus database covering the period 2015–2025 and analyzed using VOSviewer and RStudio with the bibliometrix package. The analysis employed multiple bibliometric techniques, including co-authorship, keyword co-occurrence, citation, co-citation, bibliographic coupling, and descriptive statistical analysis. The results show a substantial and exponential increase in publications, especially after 2020, with a total of 8,340 documents and an annual growth rate of 9.88%. The co-occurrence analysis reveals several major thematic clusters, ranging from technical aspects such as machine learning to pedagogical integration and the emergence of generative AI topics such as ChatGPT. Furthermore, the findings indicate a shift in research focus from technical and system-oriented approaches toward more pedagogical and normative dimensions, including issues of AI ethics and academic integrity. The co-authorship analysis shows that research collaboration remains fragmented, while global contributions are dominated by a limited number of countries. Overall, this study provides a comprehensive overview of the intellectual structure and development of AI research in higher education and highlights future research directions, particularly in interdisciplinary collaboration, ethical AI implementation, and inclusive knowledge development.