Physician AI Use Doubles in Three Years, But Trust Hinges on Performance Over Transparency

The rapid adoption of artificial intelligence in clinical settings has reached a tipping point, with more than four in five physicians now using AI professionally—more than double the rate reported just three years ago. Yet as deployment accelerates, a fundamental question remains unresolved: what actually makes doctors trust these tools enough to integrate them into patient care?

According to a 2026 American Medical Association survey of nearly 1,700 physicians, 81% now use AI in their professional work, up sharply from 2023 levels. More than three-quarters believe these tools improve their ability to care for patients. But the same survey reveals that trust remains conditional, with 88% of physicians demanding robust safety and efficacy validation before broader adoption, 86% prioritizing data privacy protections, and 85% wanting direct involvement in AI adoption decisions.

The Transparency Debate: Does Explainability Actually Matter?

Dr. Matt Sakumoto, Chief Clinical Product Officer at ambient AI provider Nabla, argued at Newsweek's AI Health Summit in September that transparency is essential for building physician confidence. "If you can't explain how it works, you're not going to trust it," he said during a panel discussion on September 23 in New York City.

But emerging evidence suggests the relationship between explainability and trust may be more complex. A systematic review of explainable AI studies found mixed results: only five of ten studies showed that explainable AI increased trust, while two found no effect, two showed mixed results, and one was inconclusive. More tellingly, a 2024 JAMA Network Open study found that AI performance—not transparency—was the factor most strongly associated with U.S. adults' preference for AI-assisted medical encounters.

Dr. Chris DeRienzo, Chief Physician Executive at the American Hospital Association, emphasized that sustained performance matters more than initial promises. "Trust is built over time through consistent, reliable outcomes," he noted at the summit.

What Physicians Actually Want From AI Vendors

The AMA survey data reveals specific priorities that extend beyond technical capabilities. Physicians want validation frameworks that continue after deployment, not just at launch. They demand involvement in procurement and evaluation processes. And they expect liability frameworks that clarify responsibility when AI-assisted decisions lead to adverse outcomes.

Theo Zanos, Head of Division of Health AI at Northwell Health, pointed to the importance of representative training data. Dr. Sophie Bakri, Medical Director of the Center for Digital Health at Mayo Clinic, stressed that clinician consultation must be embedded throughout the AI lifecycle, not treated as an afterthought. Ratnakar Lavu, Chief Digital Information Officer at Elevance Health, added that data privacy protections must be demonstrable and auditable.

Implications for the Clinical AI Market

For ambient AI vendors competing in the increasingly crowded clinical documentation market, the findings suggest that sustained real-world performance data may be a more powerful differentiator than explainability features alone. The 85% of physicians who want direct involvement in AI adoption decisions create a clear mandate for co-design and post-deployment feedback mechanisms.

As physician adoption continues to climb, the trust equation appears to favor vendors who can demonstrate consistent performance over time, involve clinicians in meaningful ways throughout the product lifecycle, and provide transparent validation data—even if the underlying algorithms remain partially opaque.

Sources

1. Doctors Are Using More AI. What Makes Them Trust It? — Newsweek