Mapping of Artificial Intelligence Applications and their Impacts on Academic Performance in Higher Education (2020–2025): A Systematic and Bibliometric Literature Review

Abstract

Artificial intelligence is increasingly used in higher education, yet evidence concerning its relationship with academic performance remains fragmented across adoption studies, predictive models and evaluations of learning outcomes. This study maps and synthesises Scopus-indexed research published between 2020 and 2025. A bibliometric analysis of 84 records was conducted using VOSviewer and Microsoft Excel, while 29 eligible studies were examined through a PRISMA-guided systematic review and structured content analysis. The review analysed publication patterns, influential contributors, AI application types, educational purposes, research designs, outcome measures and theoretical frameworks. The findings show rapid publication growth, strong reliance on quantitative and predictive designs, and increasing attention to generative AI, adaptive learning, and learning analytics. However, evidence of improved academic performance varies according to whether studies measured perceptions, prediction accuracy, correlations or direct learning outcomes. Important gaps remain in the limited use of pedagogical frameworks, including TPACK, and the scarcity of experimental, longitudinal, and geographically diverse research. The study contributes an outcome-centred synthesis that connects bibliometric patterns with the methodological and pedagogical strengths of the available evidence.