Ethics and AI Use in Medical Practice: American College of Physicians Releases Position Statement

A new position paper from the American College of Physicians aims to provide guidance for ethical use of artificial intelligence (AI) in medical practice.¹ According to the paper, AI use should be regulated by independent clinical judgment, tailored to patient needs, used to advance equitable distribution of care, and respect patient autonomy. The authors position AI as "augmented" intelligence rather than a substitute for physician judgment, building on a 2024 ACP policy paper that took the same stance.²

3 Guideposts: Relationality, Self-Governance, and Competence

Under relationality, the paper notes that physicians should use AI in ways that preserve the patient-physician relationship, arguing AI cannot be accountable to patients, appreciate individual patient values over time, or form a therapeutic alliance. The authors raise particular concern about attention: while ambient AI scribes have shown promise in reducing documentation burden and may let physicians focus more on patients, the paper cites survey data indicating nearly 60% of Americans are uncomfortable with the idea of a physician relying on AI in their own care.

Self-governance maintains physicians should use their independent clinical judgment and identify bias proactively rather than defer entirely to an AI output. Competence addresses the risk of deskilling; the paper warns that overreliance on AI tools could erode diagnostic and clinical skills, particularly among trainees and less experienced physicians.

Position Statements

The paper translates the 3 guideposts into 4 position statements: AI must support rather than weaken the patient-physician relationship; physicians and trainees must recommit to core competencies of caring while developing new competencies specific to AI; AI use should respect and promote physician-independent clinical judgment rather than replace it; and physicians, working with health systems, must actively identify and mitigate algorithmic bias, including its social and structural roots.

On transparency, the authors write that disclosure obligations should be tailored to the individual patient using clinical judgment, but that "the newness of AI and patient concerns suggest erring on the side of disclosure." They point to standardized, updated summaries of how a model works as a promising tool for supporting patient-physician conversations about AI use.

Broader Context

The paper also raises the environmental costs of AI, noting that AI-based queries consume more energy than conventional internet queries and that AI data centers in socially vulnerable communities raises additional equity concerns. The authors call for local AI governance models spanning a project's full life cycle, with obligations around transparency, disclosure, and fairness scaled to clinical risk, and for diverse teams—including patients—involved in decisions about fairness and oversight. The paper, developed by ACP's Ethics, Professionalism and Human Rights Committee (EPHRC), was approved by the ACP Board of Regents and appears in Annals of Internal Medicine.

References

1. DeCamp M, Sulmasy LS, Karches KE, et al. Ethics and professionalism in artificial intelligence and medical practice: a position paper from the American College of Physicians. Ann Intern Med. 2026.

2. Daneshvar N, Pandita D, Erickson S, et al; ACP Medical Informatics Committee and the Ethics, Professionalism and Human Rights Committee. Artificial intelligence in the provision of health care: an American College of Physicians policy position paper. Ann Intern Med. 2024 Jul;177(7):964-967.