Generative AI and Educational Assessments: A Systematic Review

Abstract

The launch of ChatGPT and the rapid proliferation of generative AI (GenAI) have brought transformative changes to education, particularly in the field of assessment. This has prompted a fundamental rethinking of traditional assessment practices, presenting both opportunities and challenges in evaluating student learning. While numerous studies have examined the use of GenAI in assessment, no systematic review has been conducted to synthesise the existing empirical evidence on this topic. Systematically reviewing 19 empirical studies published within 10 years, starting in 2014, this study assessed the current state of empirical evidence regarding GenAI in educational assessment practices and the future research directions required to advance this field. The findings were synthesised into four themes: (1) Educators’ perceptions of GenAI in assessment practices; (2) Students’ perceptions of GenAI in assessment practices; (3) Effectiveness of applying GenAI in assessment practices; and (4) Recommendations for leveraging GenAI in future assessment practices. The first three themes summarise the current empirical evidence, while the fourth theme identifies priorities for future research to guide the effective integration of GenAI into assessment practices.