Future direction of artificial intelligence in higher education assessment: a systematic review
This systematic review examines how artificial intelligence (AI), including generative AI and explainable AI (XAI), is reshaping assessment practices in higher education. Drawing on 31 empirical, conceptual, and review-based studies published between 2015 and 2025, the review identifies major shifts in the use of AI-supported grading, personalised feedback systems, AI literacy initiatives, and hybrid human–AI assessment models. While these developments offer efficiency and pedagogical innovation, they also raise critical concerns regarding academic integrity, transparency, and institutional readiness.. Evidence shows a persistent geographical imbalance: research activity remains concentrated in the Global North, where institutions often demonstrate greater policy readiness and technological capacity, while studies from Global South contexts more frequently foreground AI literacyand infrastructure.. Using the Context–Intervention–Mechanism–Outcome (CIMO) framework, the review highlights persistent gaps, including limited theoretical integration, insufficient longitudinal evidence, and underrepresentation of diverse educational contexts. Key recommendations include the development of transparent governance structures, institution-wide AI literacy initiatives, and culturally adaptive assessment frameworks. Overall, the findings underscore the need for ethical oversight, empirical rigour, and cross-cultural collaboration to ensure that AI-enabled assessment is equitable, trustworthy, and educationally sustainable in higher education.