AI in the Classroom: Observing Preclinical Students' Use of ChatGPT During Case-Based Learning at a UK Medical School

This article explores how preclinical students in a UK medical school utilise ChatGPT during their case-based learning (CBL) curriculum.

Focused ethnography was used to study 42 medical students and three clinical sciences students as they undertook seven CBL sessions over 4 weeks. In situ observations, screenshots of ChatGPT conversations and focus group data were collected and analysed using reflexive thematic analysis.

ChatGPT was used to automate concept retrieval, problem-solving and applying theory to clinical context. This occurred because students were motivated by the efficient completion of workbook questions, which could be achieved by entering them into ChatGPT. Collaborative groups were less likely to automate cognitive effort because they placed greater value on being engaged during learning and perceived socially constructing answers as more efficient than ChatGPT use. Although ChatGPT sometimes gave partial or false-but-plausible answers, students were rarely observed cross-checking it.

AI chatbot use during conventional curricular activities can result in students automating cognitive effort. This is more likely during tasks scaffolded by guiding materials and when students are motivated by external pressures rather than a desire for engaging learning. Educators should specify how and when AI chatbots should be used, promote autonomy and collaboration in student cultures and teach students about the value of cognitive effort to learning.