AI-generated multiple-choice questions in health science education: Stakeholder perspectives and implementation considerations

Abstract

Multiple-choice questions (MCQs) are widely used in health science education because they are an efficient way to evaluate knowledge from simple recall to complex clinical reasoning. The creation of high-quality MCQs, however, can be time-consuming and requires expertise in question composition. Advancements in artificial intelligence (AI), especially large language models (LLMs), offer the potential to allow for the rapid generation of high-quality, consistent, and course-specific MCQs. Here we discuss the potential benefits and drawbacks of the use of this technology in the generation of MCQs, including ensuring the accuracy and fairness of questions, along with technical, ethical, and privacy considerations. We offer practical guiding principles for the implementation of AI-generated MCQs and outline future research areas related to their impact on student learning and educational quality.