AI meets AI

A systematic review of assessment innovation and academic integrity in the age of artificial intelligence

Abstract

The rise of generative artificial intelligence (GenAI) is prompting a fundamental review of assessment design and academic integrity in higher education. This systematic review synthesises insights from 50 peer-reviewed studies and policy documents to explore how GenAI is reshaping pedagogical practice, institutional policy, and integrity norms. Evidence reveals a broad transition toward authentic, process-based, and oral assessments, alongside increased use of hybrid human–AI feedback mechanisms. A growing emphasis on AI literacy for students and staff also signals a shift in pedagogical priorities. However, the review identifies persistent tensions, including ethical ambiguity, fragmented policy responses, and over-reliance on detection tools with limited effectiveness. A distinctive contribution of this study is the Stakeholder Relevance Matrix, which highlights how emerging practices in assessment innovation, feedback design, and policy clarity unequally affect students, educators, and institutions. This matrix surfaces disparities in responsibility, vulnerability, and agency, underscoring the need for inclusive and values-driven responses. The review calls for a move from reactive enforcement toward integrity-by-design approaches that embed ethical and pedagogical principles from the outset. It concludes with forward-looking propositions to guide institutional policy and future research, offering a roadmap for building resilient, equitable, and ethically grounded assessment systems in the age of GenAI.