This Focus Session introduces participants to Generative AI for teaching clinical and communication skills in healthcare education. As traditional Standardized Patient (SP) simulations are often costly and logistically challenging, Generative AI can offer a scalable, cost-effective alternative that can be easily integrated into healthcare curricula across diverse institutional and cultural settings. As an example, we introduce the publicly and freely available AI Patient Actor (https://geiselmed.dartmouth.edu/thesen/patient-actor-app/) developed at Dartmouth. Powered by an LLM, the app simulates realistic patient interactions, providing students with the opportunity to practice clinical communication skills with an AI patient actor and receive personalized feedback. Through this experience, medical educators will learn Generative AI skills that they can apply to develop and customize their own AI-driven educational tools. Participants will engage in hands-on activities, using the app to simulate patient interactions across different clinical scenarios. This will be followed by group exercises focused on developing advanced prompts to create nuanced, realistic patient characters and clinical scenarios. We will explore strategies for adapting the Generative AI cases for equitable use, ensuring that AI-driven tools can be implemented across diverse populations. In the final section, a guided discussion will allow participants to share their experiences, challenges, and discuss best practices, with a particular emphasis on how AI-driven simulations can be tailored to various educational contexts going forward. By the end of the session, participants will have a thorough understanding of how to develop and deploy cases at their own institutions and will be equipped with the knowledge to innovate and develop customized AI-driven solutions for clinical reasoning and communication skills training.

Agenda & Methods

• 0-15 minutes: Introduction and Overview
  The session will begin with a brief introduction to the AI Patient Actor (https://geiselmed.dartmouth.edu/thesen/patient-actor-app/), its development, and its integration into medical curricula in the US and Kenya since November 2023. The facilitators will present an overview of the session’s objectives and the significance of AI-driven simulations in enhancing communication skills in medical education. This section will include a live demonstration of the app’s key features.
• 15-30 minutes: Interactive Demonstration
  Participants will engage in an interactive demonstration where they will use the AI Patient Actor in small groups. Each group will be assigned a different clinical scenario to explore, allowing them to experience firsthand how the app generates patient interactions and provides individualized feedback. Facilitators will circulate to offer guidance and answer questions.
• 30-50 minutes: Workshop on Advanced Prompting Techniques
  Participants will be divided into small groups to practice developing advanced prompts for AI-generated standardized patients. Each group will focus on different aspects, such as clinical case details, patient personality, and communication style. They will use the prompts within the app and generate their own to observe how varying inputs affect patient responses.
• 50-70 minutes: Equitable Adaptation and Implementation Strategies
  Participants will work in small groups to brainstorm adaptation strategies simulating patients from diverse backgrounds and for use in different cultural contexts and languages.
• 70-90 minutes: Group Discussion and Reflection
  Participants will share their experiences, challenges, and insights from the demonstration and prompting exercises. Facilitators will lead a discussion on best practices for creating effective AI-driven patient simulations, their promises and their limitations.