AI in Clinical Practice: Physicians Are In — Now What?

By Elena Pak, Credentialing Department, WCH

What the 2026 Doximity State of AI in Medicine Report Means for Healthcare Providers

Healthcare AI has crossed a threshold. A year ago, it was reasonable to ask whether physicians would adopt artificial intelligence tools in meaningful numbers. That question is now settled. The real questions — the ones that matter for hospital administrators, health system CIOs, and clinical operations teams — are about what comes next: how to deploy AI responsibly, how to measure its impact, and how to turn promising pilots into durable institutional practice.

The 2026 State of AI in Medicine Report, released by Doximity in March 2026 and based on survey data from 3,151 U.S. physicians across 15 specialties, provides one of the most detailed snapshots yet of where clinical AI actually stands. The findings are striking — and, for providers, they carry clear strategic implications.

From Curiosity to Routine: The Adoption Numbers

The headline statistic is remarkable: 94% of surveyed physicians reported that they are either currently using AI tools or are actively interested in doing so. More telling than the level of interest, however, is the pace of actual adoption. Daily AI usage among physicians jumped from 47% in early 2025 to 63% by January 2026 — a 16-percentage-point increase in less than a year. This is not incremental drift; it is rapid, system-wide change.

Adoption is uneven across specialties, which has direct implications for deployment strategy. Neurologists lead with a 64% adoption rate, followed by gastroenterologists (61%) and internists (60%). Family medicine physicians, while not the top adopters by percentage, are among the most intensive users: 88% of those who have adopted AI use it daily. For providers managing large primary care networks, this latter figure is especially significant — it suggests that when frontline generalists embrace AI tools, those tools become deeply embedded in workflow rather than remaining occasional conveniences.

Key Data Points at a Glance

• 94% of physicians currently use AI or are interested in doing so
• 63% daily AI usage (up from 47% in early 2025)
• 75% of AI users report reduced administrative workload
• 69% of AI users say patient care outcomes have improved
• 90% believe AI can reduce after-hours “pajama time”
• 71% cite accuracy and reliability as their top concern

Where AI Is Actually Being Used — And What That Tells Us

There is a common misconception — amplified by technology marketing — that clinical AI is primarily a diagnostic or decision-support tool. The Doximity data challenges this narrative directly. The most common AI use case among physicians is literature search, cited by 35% of respondents in the January 2026 cohort (up from 22% in April 2025). Voice-based documentation, including ambient listening and AI scribes, is the second most common application, used by 29% of physicians (up from 20%).

This is a crucial insight for provider organizations. Physicians are not primarily asking AI to help them think — they are asking it to help them work. The dominant use cases are about finding information faster and reducing the documentation burden. For health system executives, this means the most immediate ROI from AI investment lies not in complex clinical decision support, but in tools that streamline the documentation stack and surface evidence at the point of care.

The documentation burden in particular has become a defining crisis in physician wellness. Research consistently shows that for every hour of direct patient care, physicians spend roughly two additional hours on EHR data entry. This administrative overflow is a primary driver of burnout, career attrition, and what has come to be called “pajama time” — the hours physicians spend charting from home after their clinical day is done. Ninety percent of surveyed physicians believe AI has the potential to eliminate this problem; 23% say it already has.

The Trust Gap: Accuracy, Reliability, and Institutional Policy

The picture of rapid adoption sits alongside a persistent and significant concern: 71% of physicians cited accuracy and reliability as their primary worry about AI. This is not irrational resistance — it reflects the specific conditions of clinical decision-making, where an incorrect output can have consequences that extend far beyond the professional.

Nearly half of respondents (47%) reported that their hospital’s institutional policies on AI are either confusing or “still evolving.” This is a governance problem as much as a technology problem, and it may be limiting adoption in the very places where AI could have the greatest impact. When front-line clinicians cannot get clear answers about liability, documentation standards, or appropriate use cases, many will default to caution.

The trust gap is not merely attitudinal. It reflects real gaps in how AI tools are validated and communicated to clinical users. Doximity’s own response to this challenge — the development of PeerCheck, a physician-led AI review program involving over 10,000 medical experts co-chaired by Dr. Eric Topol and former U.S. Surgeon General Regina Benjamin — represents one model for addressing it. PeerCheck certifies AI-generated answers with links to reviewing physicians’ profiles, creating a chain of accountability that aligns with how clinicians already evaluate evidence.

Whether through a program like PeerCheck or through internal clinical governance structures, providers need to solve the trust problem actively. Leaving physicians to individually assess AI reliability is neither efficient nor sustainable at scale.

Strategic Implications for Provider Organizations

For health system leaders, the Doximity data suggests several concrete priorities.

1. Invest in documentation AI now. The use case with the clearest near-term ROI is ambient documentation and AI scribing. Physician time is expensive, burnout has real operational and financial consequences, and the documentation burden is well-documented and addressable. Providers that have not yet piloted ambient AI tools are falling behind peers who are already reporting measurable improvements in job satisfaction and time with patients.

2. Build governance before you scale. Nearly half of physicians say institutional AI policies are unclear. Providers should treat this as an urgent operational risk. Clear, accessible guidance on appropriate AI use — covering documentation, clinical reference, and decision support — reduces liability exposure and builds the clinician confidence necessary for sustained adoption.

3. Specialty targeting matters. Not all departments will adopt AI at the same pace or for the same purposes. Neurologists, gastroenterologists, and internists are early adopters; family medicine physicians, once converted, become high-frequency users. Investment and change management resources should reflect these patterns rather than treating the clinical workforce as a monolith.

4. Physician involvement is not optional. AI tools that are developed without physician input, or deployed without physician ownership, will face persistent resistance. Governance structures, product evaluation, and vendor selection all benefit from meaningful clinical leadership — not as a checkbox, but as a substantive design input.

***

The 2026 data marks an inflection point. Physician AI adoption has moved from early adopter territory into mainstream practice in the span of roughly 18 months. The institutions that will benefit most from this shift are those that treat AI not as a technology project, but as a clinical operations challenge — one that requires governance, workflow design, change management, and sustained attention to the question of trust.

The underlying question for providers in 2026 is no longer whether physicians will use AI. It is whether the organization has built the conditions under which AI can be used well.

Sources

Doximity, Inc. “2026 State of AI in Medicine Report.” Released March 17, 2026. businesswire.com/news/home/20260317453665/en/

HIT Consultant. “Doximity Releases 2026 Report on Physician AI Adoption Across 15 Medical Specialties.” March 18, 2026. hitconsultant.net

Fierce Healthcare. “Doximity CEO Bullish on Winning the AI Market in 2026 as It Rolls Out Physician-Led AI Review.” February 9, 2026. fiercehealthcare.com

Clinical AI Report. “Doximity Review — Clinical AI Report, 2026 Evaluation.” February 3, 2026. clinicalaireport.com/reviews/doximity

RamaOnHealthcare. “Doximity Releases 2026 Report on Physician AI Adoption Across 15 Medical Specialties.” March 18, 2026. ramaonhealthcare.com

Doximity Blog. “AI Medical Scribes: How Artificial Intelligence Is Transforming Clinical Documentation.” doximity.com

Morningstar / Business Wire. “Doximity Study Finds Physicians Rapidly Adopting AI, But Accuracy Concerns Persist.” March 17, 2026. morningstar.com