10 surprising stats about AI in healthcare from patients and physicians

Artificial intelligence (AI) is moving quickly into healthcare, but adoption isn’t just about technology. It’s about people — and how they feel about change. That’s why athenahealth partnered with the United States of Care (USofCare) to survey both physicians and patients about their views on AI in medicine.

Here are 10 surprising stats from the research — and what they tell us about the future of care.

1. 68% of physicians increased their use of AI for documentation in 2025

That’s a massive year-over-year jump, showing how quickly ambient note-taking and transcription tools are becoming mainstream*. Nearly half of physicians (48%) believe that improved clinical documentation is the biggest potential benefit of using AI in healthcare. Documentation may be the most natural entry point for AI: it solves a universal pain point and often raises record accuracy without challenging clinical authority.

2. 61% of physicians worry AI will reduce human interaction with patients

Even as adoption grows, more than half of physicians fear technology could distance them from the people they serve*. It’s a reminder that efficiency alone isn’t enough. AI must preserve and even enhance the clinician–patient relationship.

3. 77% of patients say they want to be told when AI is used in their care

Transparency isn’t a “nice-to-have;” it’s a prerequisite. Patients want clear communication about when AI is at work, whether in documentation, diagnostics, or communication. Trust begins with disclosure.

4. Only 37% of patients are comfortable with AI diagnosing conditions

Comfort levels drop sharply when AI moves into clinical decision-making*. Administrative uses like scheduling or documentation get far higher approval, but diagnosis and treatment plans remain firmly in the human domain — at least for now.

These findings underscore why equity must be central in AI design, training, and deployment.

5. 33% of Black adults believe AI could increase bias

Concerns about bias are significantly higher among communities that have historically experienced disparities in healthcare* In the surveys, 33% of Black adults expected increased bias from AI in healthcare compared to 21% of white adults. Similarly, 26% of physicians said they worry AI bias could negatively affect patient access to care. These findings underscore why equity must be central in AI design, training, and deployment.

6. 49% of patients already recognize AI is used in managing medical records

Public awareness is growing fast, with nearly half of adults identifying AI’s presence in everyday healthcare tasks*. This suggests AI is not as invisible as some may assume — and patients are paying attention.

7. 26% of patients feel optimistic about AI, but 27% are uncertain and 26% are concerned

The near-even sentiment split among patients shows how unsettled public opinion about AI in healthcare remains*. For every optimist, there’s a skeptic — making communication, education, and responsible rollouts of healthcare AI tools critical.

8. 48% of physicians believe that identifying patterns and anomalies in patient data that a physician might miss is one of the biggest potential benefits of using AI in healthcare

Doctors aren’t only using AI for administrative relief. Nearly half (48%) see it as a tool for clinical insight, highlighting opportunities for population health, diagnostics, and predictive analytics — if implemented carefully*.

9. 63% of patients believe stronger oversight of AI is needed

Patients don’t just want transparency; they want accountability*. Nearly half of patients (43%) say AI in healthcare is not well regulated, and almost two-thirds (63%) believe it requires greater oversight to ensure it’s used responsibly. This echoes physicians’ own concerns and points to a growing demand for regulators, payers, and providers to set clear guardrails. In fact,15% of physician respondents identified AI regulation as a top concern for policymakers to address — that’s up from 10% in 2024.

10. 58% of physicians fear overreliance on AI — and 52% of patients worry about improper diagnoses

Concerns converge between physicians and patients when it comes to clinical accuracy. More than half of physicians (58%) are uneasy about clinicians leaning too heavily on AI for diagnosis, while over half of patients (52%) fear the technology could lead to incorrect results*. Together, these findings highlight the shared demand for guardrails, oversight, and human judgment at the center of care.

What this research says about AI sentiment among patients and providers

Together, these 10 data points reveal a healthcare system in transition. Physicians are leaning into AI where it saves time and effort but remain more cautious about clinical use. Patients want transparency, fairness, and reassurance that AI will not compromise the human aspects of care.

The lesson is clear: adoption and acceptance are not the same. As AI in healthcare becomes more ubiquitous, patients and physicians must feel it’s trustworthy, equitable, and human-centered.

Learn more about what this research tells us about AI in healthcare.