A qualitative survey on perception of medical students on the use of large language models for educational purposes.

Abstract

Large language models (LLMs)-based chatbots use natural language processing and are a type of generative artificial intelligence (AI) that are capable of comprehending user input and generating output in various formats. They offer potential benefits in medical education. This study explored the student's feedback on the utilization of LLMs in medical education. We conducted an in-depth interview with open-ended questions with Indian medical students via telephone conversation. The recording (average time 55.28±18.04 min) was transcribed and thematically analyzed to find major themes and sub-themes. We used QDA Miner Lite v.2.0.8 (Provalis Research: Montreal, Canada) for the thematic analysis of the text. A total of 25 students from eight Indian states studying from the first to final year of studies participated in this study. Three major themes were identified about usage scenario, augmented learning, and limitation of LLMs. Students use LLMs for clarifying complex topics, searching for customized answers, solving MCQs, making simplified notes, and streamlining assignments. While they appreciated the ease of access, ready reference for getting clarity on doubts, lucid explanation of questions, and time-saving aspects of LLMs, concerns were raised regarding erroneous results, limited usage due to reliability and privacy issues, and the overreliance on chatbots for educational needs. Hence, they emphasized the need for training for the integration of LLM in medical education. In conclusion, according to students' perception, LLMs have the potential to enhance medical education. However, addressing challenges and leveraging the strengths of LLMs are crucial for optimizing their integration into medical education.