Artificial Intelligence in Graduate Medical Education

Graduate medical education (GME) must prepare our learners for a future as physicians in the world of artificial intelligence (AI). AI has been among us for decades penetrating various aspects of our lives. In recent years, generative AI has made more advances in the health care field and medical education. As medical educators there have been calls to include AI principles in training to better prepare trainees for this new AI world.¹ AI has many uses in medical education such as workload reduction, simulation, personalized education, data analysis and research, clinical decision making, task automation, content accessibility, etc.² Generative AI in medical education also comes with many challenges, including hallucinations, bias, cost, security, academic integrity and copyright issues, lack of human perspective, lack of human interaction, lack of consistent policy, and privacy protection.^3,4

As we are learning different ways to use generative AI in medical education, we must be cautious regarding over reliability on AI. AI can be helpful and harmful at the same time if not used cautiously. For example, if using generative AI to generate differential diagnosis in a patient with undifferentiated signs and symptoms, an experienced attending should be able to sort through the generated differential to know if it is appropriate for that specific patient, whereas an intern will likely not initially have that skill. This same scenario could be helpful to interns if the faculty assists them in working through the differential and could turn this into a teaching pearl. Educators need residents to learn the critical thought process of differential diagnosis prioritization based on the history and physical examination, available diagnostic workup, etc., to care for our patients as a whole person. We are at the start of our AI journey in medical education and there is a need for more study to ensure we are guiding our residents accordingly. Ethical challenges and issues with AI, data protection, liability, and ethical values and principles are main teaching recommendations.⁵ Russell, et al published competencies for the use of AI for health care professionals, identifying 6 competencies⁶:

1. Basic knowledge of AI

2. Social and ethical implications of AI

3. AI-enhanced clinical encounters

4. Evidence-based evaluation of AI-based tools

5. Workflow analysis for AI-based tools

6. Practice-based learning and improvement regarding AI-based tools

These competencies could be used to help guide medical educators in their acquisition of AI principles and use. Additionally, our STFM colleagues have created the Artificial Intelligence and Machine Learning for Primary Care (AiM-PC) curriculum to help bridge the educational gap on AI, and ACGME has a webinar, “The Generative AI Revolution: Innovations and Opportunities in Medical Education.”

Regardless, AI is already here and not going anywhere. Every day we see more use of AI in the health care field and in medical education. While best practices are being studied and developed nationally and within our specialty, at minimum ensuring that your program or institution has guidelines or policies on use of generative AI is imperative. As GME educators and innovators, we must learn safe use of AI in the medical field and create a pathway for safe and effective use of AI for our future generations.