Artificial intelligence in educational assessment: A bibliometric mapping of research trends (2020–2026)

Abstract

Background/purpose.  Artificial intelligence (AI) has increasingly transformed educational assessment through automated grading systems, adaptive learning technologies, and personalized feedback mechanisms. The rapid development of generative AI and digital learning environments has significantly increased scholarly interest in AI-driven educational assessment. Therefore, this study aimed to examine the global research trends, thematic structures, and intellectual development of artificial intelligence in educational assessment research published between 2020 and 2026.

Materials/methods. This study employed a bibliometric research design using data obtained from Google Scholar through the Publish or Perish application. A total of 500 publications related to AI in educational assessment were collected and analysed. The data were exported in RIS and CSV formats and analysed using VOSviewer and Microsoft Excel to identify publication trends, keyword co-occurrence networks, thematic clusters, and emerging research patterns.

Results. The findings revealed a substantial increase in publications concerning AI-driven educational assessment, particularly after 2023. The bibliometric mapping identified five major thematic clusters dominated by terms such as artificial intelligence, adaptive learning, automated grading, personalized feedback, and machine learning. The overlay visualization further demonstrated that recent studies have increasingly focused on generative AI, ChatGPT, adaptive learning technologies, and real-time feedback systems as emerging research trends.

Conclusion. Research on AI in educational assessment has become a rapidly growing, multidisciplinary field that integrates technological, pedagogical, and data-driven perspectives. Although AI-driven assessment demonstrates considerable potential in improving efficiency and personalization, ethical concerns related to fairness, transparency, and human oversight remain important issues for future educational research and practice