The trend: 81% of physicians report that their healthcare organization is using AI in some capacity, according to a February 2026 survey from the American Medical Association (AMA) of nearly 1,700 clinicians across various specialties, practice settings, and career stages.

• AI adoption among physicians has surged, with the 72% who reported at least one known use case up sharply from 48% in the AMA’s 2024 survey and 38% in 2023.

Unpacking the trend: Doctors are expanding AI use across their workflows and growing more optimistic about its impact on patient care.

• The average number of AI applications used by physicians rose from 1.1 in 2023 to 2.3 this year, with AI use increasing across most physician workflows surveyed.
• Using AI to summarize medical research and standards of care is the top use case (39%), up significantly from 10% in 2024.
• Physicians expect AI to be most valuable in improving workflow efficiency, reducing cognitive overload, and supporting diagnostic decision-making.
• 76% of clinicians say AI improves their ability to care for patients, up from 69% in 2024.

However, physicians do report concerns about AI in clinical settings, with deskilling cited as the top risk.

• Nearly 9 in 10 (88%) of doctors express at least mild apprehension about health AI-related skill loss.
• 70% are “very” or “somewhat” concerned about the loss of physician skills among the medical students and residents being trained today.

Why it matters: Hospitals and health systems are fueling the healthcare sector’s surge in AI investment, implementing commercial AI solutions at more than twice the rate (2.2x) of the broader US economy, according to an October 2025 report from Menlo Ventures. Provider organizations account for about three-quarters of healthcare’s commercial AI investment total, while nearly all health system C-suite executives report plans to implement AI solutions if they haven’t already done so, per an August 2025 Sage Growth survey.

Implications for healthcare organizations: Hospital and health system leaders must ensure that doctors trust AI enough to use it in daily workflows, yet don’t fear it will become so capable that it deskills them or even replaces them. That’s a difficult balance to strike, especially as clinicians expand their use of AI and recognize its benefits.

Provider organizations will need to increase training that emphasizes human oversight, requiring physicians to review AI-supported outputs and justify their decisions. At the same time, they must distinguish between tasks where some deskilling is acceptable and those where maintaining physician expertise and critical thinking is essential.