Abstract
The integration of Artificial Intelligence (AI) in medical education is rapidly transforming assessment practices, offering unprecedented opportunities to enhance student evaluation, feedback, and learning pathways. However, despite the potential, a comprehensive understanding of these opportunities and their interdependencies has been lacking. This study provides a critical review of the literature on AI's role in medical education assessment, categorising 22 identified opportunities into seven major "mega-opportunities" that address various aspects of student assessment. Through the application of Interpretive Structural Modelling (ISM), the cause-effect interdependencies among these mega-opportunities were explored, revealing a complex web of relationships that guide their effective implementation. The findings highlight the central role of "Automated Feedback and Evaluation" and "Data-Driven Analytics and Curriculum Improvement" as foundational drivers, with far-reaching impacts on other areas like "Simulation-Based Assessment" and "Longitudinal Assessment and Development." This paper culminates in the proposal of aresearch roadmap that highlights the priority of addressing different mega-opportunities in AI and assessment, offering practical guidelines for medical researchers, educators, institutions, and policymakers to adopt AI-driven assessment strategies. Future research avenues are identified to explore the real-world application and impact of these AI-driven innovations, focusing on longitudinal studies and educational equity. The findings underscore the need for continued research to refine the model proposed by this study and adapt it to diverse educational environments.