This guidance applies to residents, fellows, and supervising faculty involved in graduate medical education and relates to the appropriate use of AI tools in the clinical training and patient care environments. AI tools may support learning and clinical workflows, but they must be used in ways that prioritize patient safety, uphold professional standards, and reflect supervised, competency-based practice. AI supports clinical education and decision-making but does not replace clinical judgment or supervision. Residents and fellows who also engage in research, teaching, or administrative work should consult the appropriate role-based guidance.
-
AI tools may be used to support clinical education and practice by assisting with learning and clinical reasoning exercises; supporting access to point-of-care resources; and helping summarize clinical information for review. AI tools are intended to support, not replace, clinical judgment, professional expertise, or supervision. Final clinical decisions must always be made by qualified clinicians using appropriate evidence, experience, and patient-specific context.
-
AI use by residents and fellows occurs within a supervised educational framework. Supervision ensures appropriate integration of AI into clinical learning and practice. Residents and fellows:
- Should use AI tools under the guidance of supervising faculty when AI contributes to clinical decision-making.
- Must follow program- and department-specific guidance regarding AI use.
- May be subject to competency-based expectations before independent use of certain AI tools.
-
Clinical judgment remains a human responsibility. Residents and fellows are accountable for understanding and critically evaluating AI-generated information; integrating AI output with clinical knowledge and patient context; and escalating questions or concerns to supervising clinicians. AI tools must not be used as substitutes for professional judgment or independent clinical reasoning.
-
Patient privacy and confidentiality are paramount. Residents and fellows must not enter protected health information (PHI) or other sensitive patient data into unvetted or unapproved AI tools. Only AI tools that have been vetted and approved for clinical use may be used with patient information, and only within approved workflows.
-
In some contexts, the use of AI tools may need to be documented in the medical record. Patients may need to be informed of AI use, when applicable. These requirements may vary by program, department, or clinical setting. Residents and fellows should follow applicable clinical guidance.
-
Transparency and disclosure of AI use are required when AI tools contribute meaningfully to clinical documentation, decision-support activities, or educational materials. Disclosure supports trust, patient understanding, and professional accountability.
-
Some AI tools may require demonstration of AI-related competencies before independent use. Competency expectations may include understanding tool capabilities and limitations, appropriate interpretation of AI-generated information, and awareness of risks, bias, and ethical considerations. Program leadership may define additional requirements based on clinical context.
-
Residents and fellows should be attentive to potential bias in AI-generated outputs. AI tools may reflect patterns that do not account for individual patient circumstances or diverse populations. Human review is essential to ensure equitable and patient-centered care.
-
Residents and fellows are encouraged to consult program-specific guidance, department leadership, and institutional AI guidance on Approved Tools and Governance, Data Protection and Safe Use, and Transparency and Disclosure.
When You Are Unsure
If you are uncertain whether AI use is appropriate in a clinical or educational context, pause before proceeding. Follow institutional guidance for when you are unsure and consider consultation with appropriate leadership, such as supervising faculty or program leadership. Requests for assistance can also be sent to Support+AI@wmed.edu.