Proof of Concept of ChatGPT as a Virtual Tutor

Am J Pharm Educ. 2024 Nov 14;88(12):101333. doi: 10.1016/j.ajpe.2024.101333. Online ahead of print.

Abstract

Objective: This study aimed to evaluate the feasibility and perceived value of an instructor-developed ChatGPT-based virtual tutor for third-year Doctor of Pharmacy students in a pharmacy management course. The objectives were to assess faculty time required to develop the tool and analyze student engagement and perception of its educational value.

Methods: The virtual tutor was developed by the instructor and consisted of custom ChatGPT prompts deployed via the course's learning management system. Development metrics were documented, including time invested and iterations needed for prompt refinement. A after-use questionnaire evaluated student uptake and experience, which included frequency of use, perceived value for learning and exam preparation, and overall satisfaction.

Results: The development of the virtual tutor required a total of 119 min, demonstrating a feasible time investment. Ninety-two of 130 students completed the questionnaire for a 70% response rate. Sixty-one students (66%) engaged with the tutor, with varying degrees of use. Most students (82%) reported increased confidence for exams, and 99% acknowledged enhanced understanding of the course material through the tutor.

Conclusion: The virtual tutor proved to be a feasible educational tool, with considerable student uptake and satisfaction. It offered a time-efficient alternative for educational support, potentially reducing the need for traditional face-to-face tutoring. The study highlights the effectiveness and practicality of ChatGPT as a virtual tutor in pharmacy education, but further research should explore its application across different content areas as well as potential unintended consequences.

Keywords: Artificial intelligence; ChatGPT; Virtual tutor.