Co-Designing a Chatbot for Culturally Competent Clinical Communication: Experience and Reflections

TL;DR

This paper explores the development and pilot testing of an AI-driven chatbot designed to enhance culturally competent clinical communication training for medical students, highlighting its potential benefits and current limitations.

Contribution

It introduces a novel AI chatbot tailored for cultural competence training in medical education and reflects on its initial implementation and insights.

Findings

Chatbot provided useful reflection opportunities for students.

Challenges included addressing systemic issues and nonverbal cues.

Limitations involved over-agreeableness and lack of nonverbal communication.

Abstract

Clinical communication skills are essential for preparing healthcare professionals to provide equitable care across cultures. However, traditional training with simulated patients can be resource intensive and difficult to scale, especially in under-resourced settings. In this project, we explore the use of an AI-driven chatbot to support culturally competent communication training for medical students. The chatbot was designed to simulate realistic patient conversations and provide structured feedback based on the ACT Cultural Competence model. We piloted the chatbot with a small group of third-year medical students at a UK medical school in 2024. Although we did not follow a formal experimental design, our experience suggests that the chatbot offered useful opportunities for students to reflect on their communication, particularly around empathy and interpersonal understanding. More challenging areas included addressing systemic issues and historical context. Although this early version of the chatbot helped surface some interesting patterns, limitations were also clear, such as the absence of nonverbal cues and the tendency for virtual patients to be overly agreeable. In general, this reflection highlights both the potential and the current limitations of AI tools in communication training. More work is needed to better understand their impact and improve the learning experience.