Background: Residency programs across the country face mounting challenges in sustaining high-quality didactics amid rising service needs, staffing pressures, and increasing clinical complexity. Traditional lecture-based models often fail to engage learners who benefit from dynamic, case-based, and interactive educational formats. Simulation is known to be one of the most effective modalities for improving diagnostic reasoning and crisis management skills, but access is limited by staffing, scheduling, equipment, and dedicated lab space. To bridge this gap, we sought to leverage rapidly evolving artificial intelligence (AI) tools to bring simulation-style learning directly into daily resident workflows.
Purpose: To design, implement, and pilot a flexible, low-cost, AI-enabled teaching model that enhances engagement, expands access to simulation-based education, and improves residents’ comfort with clinical reasoning and rapid response decision-making during routine teaching sessions and inpatient work.
Description: We developed the “AI Sim Lab,” an educational framework utilizing commercially available AI platforms to generate realistic, interactive clinical cases that mimic real-time simulation encounters. These AI-driven cases were deployed during morning report, noon conference, night-float teaching, and unplanned downtime on inpatient services.Using structured prompts and customizable templates, chief residents created scenarios across multiple disciplines, including shock states, electrolyte emergencies, respiratory failure, sepsis, and neurologic deterioration, tailored to PGY level and mapped to ACGME milestones. Each case incorporated dynamic branching (changing vitals, labs, and imaging based on resident decisions), micro-debriefs, and reflective questions.Implementation required no financial investment and minimal preparation time(< 5 minutes). Faculty ensured content validity and provided oversight. Residents consistently demonstrated increased participation, with notably stronger engagement from quieter learners and interns early in training. Over several months, residents reported improved confidence in handling acute inpatient scenarios, and many began adapting AI tools independently for peer teaching. Faculty noted smoother team communication during actual rapid responses, suggesting improved cognitive readiness.
Conclusions: AI-powered simulations offer a scalable, accessible, and resource-efficient solution to revitalize didactics in hospital-based training programs. By blending modern technology with traditional educational principles, the AI Sim Lab expands simulation access, strengthens diagnostic reasoning, and enhances learner engagement within everyday clinical workflows. This innovation is adaptable to diverse institutional contexts, requires no additional funding, and empowers chief residents and faculty to modernize medical education for the digital era.