Entrustment and EPAs for Artificial Intelligence (AI): A Framework to Safeguard the Use of AI in Health Professions Education
- 14 November 2024
- journal article
- scholarly perspective
- Published by Oxford University Press (OUP) in Academic Medicine
- Vol. 100 (3) , 264-272
Abstract
In this article, the authors propose a repurposing of the concept of entrustment to help guide the use of artificial intelligence (AI) in health professions education (HPE). Entrustment can help identify and mitigate the risks of incorporating generative AI tools with limited transparency about their accuracy, source material, and disclosure of bias into HPE practice. With AI’s growing role in education-related activities, like automated medical school application screening and feedback quality and content appraisal, there is a critical need for a trust-based approach to ensure these technologies are beneficial and safe. Drawing parallels with HPE’s entrustment concept, which assesses a trainee’s readiness to perform clinical tasks—or entrustable professional activities—the authors propose assessing the trustworthiness of AI tools to perform an HPE-related task across 3 characteristics: ability (competence to perform tasks accurately), integrity (transparency and honesty), and benevolence (alignment with ethical principles). The authors draw on existing theories of entrustment decision-making to envision a structured way to decide on AI’s role and level of engagement in HPE-related tasks, including proposing an AI-specific entrustment scale. Identifying tasks that AI could be entrusted with provides a focus around which considerations of trustworthiness and entrustment decision-making may be synthesized, making explicit the risks associated with AI use and identifying strategies to mitigate these risks. Responsible, trustworthy, and ethical use of AI requires health professions educators to develop safeguards for using it in teaching, learning, and practice—guardrails that can be operationalized via applying the entrustment concept to AI. Without such safeguards, HPE practice stands to be shaped by the oncoming wave of AI innovations tied to commercial motivations, rather than modeled after HPE principles—principles rooted in the trust and transparency that are built together with trainees and patients.Keywords
This publication has 67 references indexed in Scilit:
- Clinical Reasoning of a Generative Artificial Intelligence Model Compared With PhysiciansJAMA Internal Medicine, 2024
- The Limits of Clinician Vigilance as an AI Safety BulwarkJAMA, 2024
- Assessing supervisor versus trainee viewpoints of entrustment through cognitive and affective lenses: an artificial intelligence investigation of bias in feedbackAdvances in Health Sciences Education, 2024
- AI’s Threat to the Medical ProfessionJAMA, 2024
- Will Generative Artificial Intelligence Deliver on Its Promise in Health Care?JAMA, 2024
- Trust in artificial intelligence: From a Foundational Trust Framework to emerging research opportunitiesElectronic Markets, 2022
- The role of data science and machine learning in Health Professions Education: practical applications, theoretical contributions, and epistemic beliefsAdvances in Health Sciences Education, 2020
- Differences in Narrative Language in Evaluations of Medical Students by Gender and Under-represented Minority StatusJournal of General Internal Medicine, 2019
- Hedging to save face: a linguistic analysis of written comments on in-training evaluation reportsAdvances in Health Sciences Education, 2015
- Clinical Oversight: Conceptualizing the Relationship Between Supervision and SafetyJournal of General Internal Medicine, 2007