Mapping the Research Landscape of Transparent AI in University Assessment: A Bibliometric Investigation

This study presents a systematic bibliometric investigation of AI transparency research in university assessment contexts. Following GLOBAL recommendations (Ng et al., 2024), we examined 72 peer-reviewed publications from Scopus (2019-2025) using performance metrics and science mapping techniques. Findings reveal exponential growth from single publications in 2019 to 26 documents in 2024 (R² = 0.7666, p = 0.0098). The domain generated 655 citations achieving h-index of 13. China leads in output (8 documents) while Sweden demonstrates highest citation efficiency (25.50 citations per document). Science mapping identifies four primary clusters: technical transparency methodologies, educational analytics frameworks, machine learning applications, and performance prediction systems. Co-citation analysis establishes Adadi and Berrada’s XAI survey (2018) as the foundational framework (11 citations, 24 total link strength). Temporal evolution shows progression from basic concepts toward practical implementations, reflecting regulatory compliance following GDPR and EU AI Act. International collaboration reveals South-South partnerships and high-impact contributions from countries with strong data protection frameworks. These patterns provide evidence for an emerging interdisciplinary domain addressing AI accountability in higher education, offering insights for researchers, practitioners, and policymakers.