Summary
Peer-reviewed SGIM position statement that describes capabilities and limitations of generative AI systems in medicine and sets principles and practical recommendations for their use in clinical care, education, and research. Addresses human oversight, documentation, validation, data security, and governance of generative AI tools.
Healthcare Implications
Provides frontline clinicians, informatics teams, and health systems with concrete guidance on when and how to use generative AI (for example, for note drafting, patient communication, or clinical reasoning support) while preserving clinician accountability. Emphasizes safeguards to reduce hallucinations and bias, protect patient privacy, and ensure rigorous evaluation and ongoing monitoring of generative AI deployed in healthcare settings.