Workflow for Developing Practice Questions Using ChatGPT

This workflow outlines a step-by-step method for medical students and educators to efficiently create high-quality practice questions using ChatGPT to support AI-enhanced learning in medical education. It addresses key challenges such as faculty time constraints and limited access to customized study materials. By enabling students and educators to create content aligned with course and learning objectives, this workflow boosts engagement, promotes active learning, and improves exam performance. Applicable across institutions, this approach reduces faculty workload, while empowering learners to take ownership of their study process—offering a scalable, low-burden solution to enhance academic success in demanding medical curricula. 

Streamlining the question development process saves significant time while promoting active recall and strengthening long-term retention.1  To assess its effectiveness in practice, this workflow was implemented at the University of Nebraska Medical Center. A total of 211 practice questions were distributed across five exams within two system-based blocks. Unpublished data demonstrates improvements in both exam performance and student participation over time. Further, appropriate discriminatory power was confirmed in accordance with methods outlined by Laupichler et al.2  

Explore a sample quiz created using this workflow.  

Note: ChatGPT (GPT-4o) was used from 6/20/25 to 8/8/25 to test the workflow described in this resource. 

Last updated October 2025 

Authors 

Jonathan Paradis, Medical Student, University of Nebraska Medical Center 

Currey Zalman, Resident, University of Nebraska Medical Center

Geoffrey Talmon, Professor, University of Nebraska Medical Center

Kari Nelson, Professor, University of Nebraska Medical Center