Mapping the Knowledge Structure and Emerging Trends of Artificial Intelligence-Enabled Digital Transformation in Higher Education: A Bibliometric Analysis (2015–2025)

Abstract

Over the past decade, Artificial Intelligence (AI) has shifted from a peripheral technological novelty to a core catalyst driving Digital Transformation (DX) across global Higher Education Institutions (HEIs)[cite: 1]. Despite an explosive increase in relevant publications, existing literature lacks a granular, longitudinally consistent mapping of how AI-enabled educational paradigms have structurally evolved.This bibliometric study addresses this critical gap by systematically retrieving and analyzing 2,845 peer-reviewed articles published between 2015 and 2025 from the Web of Science Core Collection and Scopus databases. Utilizing VOSviewer, CiteSpace, and thebibliometrixR package, we perform multi-dimensional mapping including co-citation analysis, keyword co-occurrence clusters, thematic trajectory modeling, and institutional collaboration network profiling.Our findings reveal a three-stage evolutionary trajectory: Foundation Phase (2015–2018): Centered on early intelligent tutoring systems and algorithmic learning analytics;Expansion Phase (2019–2022): Driven by remote learning demands during global disruptions, focusing on automated assessment and predictive retention modeling; and Generative Transformation Phase (2023–2025): Dominated by Large Language Models (LLMs), multimodal AI, hyper-personalized learning pathways, and ethical governance frameworks.Furthermore, we identify five core thematic clusters defining current academic research and highlight emerging trends, notably agentic AI assistants, dynamic curriculum adaptation, and equitable AI policy formation. This study synthesizes a decade of research to present an integrative conceptual framework, offering strategic guidance for academic researchers, educational leaders, and policy-makers navigating the future of AI-driven higher education.