Physician AI Adoption Has More Than Doubled Since 2023: What the AMA's Latest Survey Reveals
A newly released American Medical Association survey finds that 81% of physicians now report using augmented intelligence (AI) tools in their professional practice—more than double the rate recorded when the AMA first surveyed doctors on the technology in 2023. The finding, drawn from nearly 1,700 respondents across specialties, practice settings, and career stages, offers one of the most detailed pictures yet of how quickly AI has moved from novelty to routine clinical infrastructure.
A Rapid Shift in Physician Behavior
The AMA's "2026 Physician Survey on Augmented Intelligence" was conducted by the organization's newly established Center for Digital Health and AI, launched last fall to help physicians shape how these technologies are integrated into medicine. The scale of adoption growth is notable: physician use of health AI to summarize medical research and stay current on standards of care rose by 33 percentage points compared with the 2023 survey, making it the single most common application, cited by 39% of respondents.
Beyond literature review, physicians reported using AI for a range of documentation-adjacent tasks: 30% for creating discharge instructions, care plans, or progress notes; 28% for documenting billing codes, medical charts, or visit notes; and another 28% for generating chart summaries. Smaller but meaningful shares reported using AI for draft responses to patient portal messages (19%), translation services (18%), and assistive diagnosis (17%).
Confidence in the technology has grown alongside usage. More than three-quarters of physicians now believe AI improves their ability to care for patients, up from 65% in 2023, with the greatest expected gains concentrated in diagnostic accuracy and work efficiency. Seven in ten physicians see AI as a tool capable of automating tasks that contribute to burnout, and 76% say it can meaningfully help with patient care.
Optimism Paired With Persistent Caution
Despite rising adoption, the survey found physician sentiment toward AI remains far from uncritical. Roughly 40% of respondents described themselves as equally excited and concerned about the technology, with patient privacy and the integrity of the patient-physician relationship cited as top concerns. AMA CEO and Executive Vice President John Whyte, MD, MPH, framed the moment as one requiring deliberate stewardship rather than passive acceptance.
"AI has quickly become part of everyday medical practice," Whyte said, adding that physicians see genuine promise in its capacity to support clinical decisions and reduce administrative burden.
That caution becomes more pronounced when the survey turns to patient-facing use. While physicians were generally supportive of patients using AI for general health and medication questions, most expressed wariness about applications requiring clinical judgment. Nearly half of respondents strongly opposed patients using AI to interpret their own radiology or pathology results—a finding that underscores where physicians believe the line between convenience and risk should be drawn.
Skill Loss and Governance Concerns
Perhaps the survey's most striking finding involves workforce development rather than patient care directly. Eighty-eight percent of physicians reported at least some concern about AI-related skill loss, and 70% said they were "very" or "somewhat" concerned specifically about skill erosion among medical students and residents currently in training. This suggests that even as physicians embrace AI's efficiency gains, many worry about the long-term effects of algorithmic reliance on clinical reasoning development.
Physicians were similarly clear about what they need before broader adoption feels safe. Data privacy protections were rated important by 86% of respondents, and robust safety and efficacy validation by 88%. Clear liability frameworks ranked as the top regulatory priority for building trust and expanding adoption. Notably, 85% of physicians said they want to be consulted or directly involved in decisions about AI adoption within their own practices or institutions—a strong signal that top-down technology mandates, without physician input, are likely to encounter resistance.
Industry Context and Momentum
The survey's findings align with broader industry movement. Last year, the JAMA Network signed a strategic content agreement with OpenEvidence, a platform the company estimates is used daily by 40% of U.S. doctors. Under the agreement, published content from JAMA, JAMA Network Open, and the eleven JAMA specialty journals now informs answers delivered through the platform—illustrating how quickly credentialed medical publishing and AI query tools are converging.
Implications for Practice Leadership
For private practice physicians and healthcare administrators, the survey's message is twofold. First, AI has crossed a functional adoption threshold: with four in five physicians already using it, the operative question is no longer whether to integrate these tools but how to govern them responsibly. Second, physician trust is conditional and specific—strongest around efficiency and documentation tasks, and notably more guarded around diagnostic interpretation and skill development in trainees.
Practice leaders evaluating AI governance frameworks should take particular note of the liability and privacy priorities physicians flagged, as well as the strong preference for inclusion in adoption decisions. As the AMA's STEPS Forward® collection and similar resources continue to evolve, institutions that involve physicians directly in AI governance—rather than treating adoption as a purely administrative decision—appear best positioned to translate rising physician comfort into safe, durable clinical workflows.
