Generative artificial intelligence for teaching and assessment in health professions education: A scoping review

AIM: To review the literature on generative artificial intelligence for teaching and assessment in health professions education BACKGROUND: Advancements in generative artificial intelligence (GenAI), such as ChatGPT, offer new possibilities for health professions education. These technologies offer potential benefits in teaching and assessment, including personalised learning and automated resource generation. Despite its potential, concerns about accuracy, ethics and reliability remain. This scoping review examines GenAI's implementation, benefits and challenges in teaching and assessment across health professions education. DESIGN: Scoping review. METHODS: Following Arksey and O'Malley's five-stage framework, with refinements based on the Joanna Briggs Institute (JBI) methodology, Medline, CINAHL and Web of Science Core Collection were searched for peer-reviewed studies published between January 2019 and June 2024. Studies were screened independently by two reviewers and data extraction performed systematically to ensure consistency. RESULTS: Studies (n = 5826) were assessed for eligibility, with 23 meeting the inclusion criteria. All included studies were published in 2023 and 2024. The primary applications of GenAI were in learning resource development and assessment, with reported benefits such as time savings, personalised learning and reduced resource use. Challenges included accuracy concerns, inconsistent outputs, technical limitations, algorithmic bias and risks to academic integrity. CONCLUSIONS: This scoping review provides an overview of how GenAI is being integrated into health professions education. While the technology offers opportunities to enhance teaching and assessment, its implementation requires consideration of reliability, ethical concerns and educator preparedness. This review is the first to examine GenAI implementation across multiple AHPRA-regulated health professions and proposes a practical framework (AI HPE checklist) to guide responsible use.