GPT-4 versus human authors in clinically complex MCQ creation: A blinded analysis of item quality
GPT-4 can produce MCQs testing clinically complex concepts for medical assessment. While the structural quality of AI-generated MCQs is comparable to experts overall, human oversight is necessary to ensure content validity and optimize item quality.
Authors
- Hannah Pham (ORCID: https://orcid.org/0000-0003-4082-2677)
- Edward Palmer (ORCID: https://orcid.org/0000-0001-9654-5213)
- Stefan Court‐Kowalski (ORCID: https://orcid.org/0000-0003-3859-6025)
- Daniel Lee (ORCID: https://orcid.org/0000-0002-4007-8692)
- Toby Zerner
- Peter G. Devitt
Institutions
- The University of Adelaide (AU)
Publication Details
- Journal
- Medical Teacher
- Published
- 2025-05-29
- DOI
- https://doi.org/10.1080/0142159x.2025.2505122
- Citations
- 8
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 5.23