American College of Clinical Pharmacy
      Search      Cart
         
  Poster Hall

Mon-32 - Engineering a Custom GPT to Assess Pre-Class Reading Comprehension in Pharmacy Education: The Development of “Professor Pill”

Scientific Poster Session III - Students Research-in-Progress

Students Research in Progress
  Monday, October 14, 2024
  01:00 PM–02:30 PM

Abstract

Introduction: The integration of Artificial Intelligence (AI) into post-secondary education is rapidly transforming the learning experience. Concerns regarding plagiarism and circumvention of critical thinking skills may lead to hesitancy in implementation; however, increased exposure through intentional competency development, such as appropriate prompt engineering and realization of output limitations (i.e., biases), offers exciting opportunities to enhance critical thinking skills and increase learner engagement. Custom Generative Pre-trained Transformers (GPTs) can be trained to meet specific curricular needs, enhancing efficiency in assignment creation and completion. By providing interactive tools for pre-class assignments, educators can ensure foundational knowledge acquirement, setting the stage for deeper learning and more meaningful classroom discussions. This project aims to showcase how the development, implementation, and evaluation of a custom GPT can complement and elevate pharmacy education.

Research Question or Hypothesis: (1) What is the process of creating a custom GPT for educational purposes? (2) What is the impact on class preparedness?

Study Design: This qualitative study explores students’ experiences and perceptions using a custom GPT for the promotion of critical thinking skills and pre-work completion in a pharmacy elective course. Data collection includes GPT transcripts, student reflections, and research reflections. Data will be analyzed thematically and triangulated through multiple data sources.

Methods: A custom GPT, “Professor Pill,” was engineered to assess the completion and comprehension of a pre-reading assignment. ChatGPT Plus with GPT-4o was utilized with system prompt engineering defining role, style, purpose, instructions, and guardrails. The knowledge base was limited to the assigned reading. Internal testing and re-training were performed to ensure optimal functionality and accuracy. A standardized user prompt was developed for interaction with “Professor Pill”. Upon completion of the interaction, students will be instructed to download the transcript for submission and asked to critically evaluate and reflect on the bot's responses to identify errors.

Results: Pending project completion.

Conclusion: Pending project completion.

Presenting Author

Macaleigh Mancuso Pharm.D. Candidate
Auburn University Harrison College of Pharmacy

Authors

Elizabeth Covington Pharm.D., BCIDP
Auburn University Harrison School of Pharmacy

Lindsey Moseley Pharm.D., Ph.D., M.Ed.
Auburn University Harrison College of Pharmacy

Courtney Watts Pharm.D., M.S., BCPS, BCOP
Auburn University Harrison School of Pharmacy

Cookies

This website uses cookies to help ACCP provide you with the best user experience. If you continue to use our services, ACCP will assume that you agree to the use of such cookies. You can find out how to update your settings by referring to ACCP’s Policy on Cookies.