Training GPT as a Standardized Patient

Resum

The integration of artificial intelligence (AI) into learning environments poses a challenge in advancing towards more efficient interactive methodologies. The use of AI-based learning assistants, especially generative language models like OpenAI’s GPT, can expand the scope of methodologies such as clinical simulation by generating interactions where AI assumes the role of a standardized patient. Clinical simulation recreates, substitutes, and/or extends real experiences through guided experiences that evoke or replicate substantial aspects of the real professional context in a fully interactive manner. The standardized patient is an actor trained to perform predefined responses based on the students’ behaviour and performance. With appropriate AI training, focused on instruction and adaptation to different patient profiles based on their health-disease processes, it is possible to design and implement clinical simulation scenarios where students interact with it. The authenticity of AI allows achieving a high degree of fidelity, and its scope surpasses the limits of synchronous in-person demand of a standardized actor, exponentially multiplying the capacity to generate simulated learning environments. This chapter outlines the keys to integrating AI as a standardized patient into simulated learning experiences.