Abstract
This paper presents a feasibility study investigating the use of a large language model-driven virtual patient voicebot for medical communication training. The voicebot simulates three distinct virtual patients based on standardized patient role descriptions commonly used in medical education. The study aims to evaluate the feasibility of using LLMs for simulating realistic amnesis interviews. Feedback from medical professionals was collected to assess feasibility in simulating human-like interactions and educational relevance. Initial results indicate promising potential for integrating LLM-driven VPs into medical training, offering realistic, scalable patient interactions. The study provides a foundation for further research on the use of conversational agents as virtual patients in healthcare education
Recommended Citation
Göbel, Johannes; Betke, Hans; and Thews, Oliver, "Virtual Patient Encounters: Assessing the Feasibility of
LLM-Driven Voicebots in Medical Communication Training" (2025). ACIS 2025 Proceedings. 13.
https://aisel.aisnet.org/acis2025/13