When physicians and patients view AI in different ways

Nationwide survey results from doctors and patients reveal a mixture of optimism and skepticism about AI in healthcare, depending on how it’s used

Artificial intelligence (AI) is no longer a futuristic concept in healthcare. From analyzing medical images to supporting clinical documentation, it has become part of the daily conversation about how care is delivered. But while adoption is increasing, the success of AI in healthcare will not depend solely on what the technology can do. It will hinge on how the people who use it — physicians and patients — perceive its value, risks, and impact.

That’s the question at the heart of new collaborative research from athenahealth and the United States of Care (USofC). By combining athenahealth’s 2025 Physician Sentiment Survey with USofC’s national poll of patients, the report offers a rare, side-by-side view of how doctors and patients are experiencing the rise of AI in medicine. The findings highlight a fascinating dynamic: patients and physicians see overlapping potential in AI, but they differ in what they find acceptable, trustworthy, and useful.

Physicians: Pragmatic adoption, lingering reservations

Physicians are becoming increasingly familiar with AI in healthcare. In fact, 61% are “somewhat or very familiar” with AI applications, and adoption has surged in the past year. Most notably, the use of AI for clinical documentation rose by 68% between 2024 and 2025, as more clinicians turn to ambient note-taking and transcription tools to reduce after-hours charting¹.

Physicians also see real benefits in AI’s ability to identify patterns (48%), improve documentation (48%), and reduce administrative burden (46%)². These uses are not about replacing clinical judgment but rather supporting it — enabling doctors to spend more time with patients and less time wrestling with paperwork.

Yet, enthusiasm has limits. Significant numbers of physicians worry about the loss of human interaction in care (61%), overreliance on AI for diagnosis (58%), and the risk of improper diagnosis (53%)³. Comfort is highest for AI used in administrative support, patient education, and documentation — areas that complement, rather than compete with, physician expertise. Far fewer doctors support using AI to guide treatment decisions or determine diagnoses.

Patients: Cautious optimism, context matters

On the patient side, awareness of AI in healthcare is rising. Nearly half of adults recognize its use in managing medical records (49%), imaging (44%), and prior authorizations (43%)⁴. Younger generations — Millennials and Gen Z — are especially aware, reflecting broader exposure to technology.

But awareness does not equate to comfort. Patients report a mixed outlook: roughly equal shares identify as optimistic (26%), uncertain (27%), or concerned (26%), while only 17% feel comfortable and 15% are excited⁵. Context matters enormously. About half of patients are comfortable with AI in simple, administrative tasks such as recording notes (49%), analyzing data (49%), or communicating test results (47%). Comfort drops when it comes to AI in high-stakes clinical decisions: 41% support its use in treatment planning, 37% in diagnosis, and 33% in performing surgery⁶.

Bias is another concern patients have about AI use in healthcare. Black adults (33%) are more likely than White adults (21%) to anticipate increased bias from AI tools⁷. Patients want reassurance that AI systems will be trained on diverse data and actively monitored for fairness. Of Hispanic adults, 30% reported greater trust in AI when organizations proactively addressed bias and promoted fairness.

Above all, patients want transparency. A full 77% believe they should be informed when AI is used in their care, and 63% say stronger oversight is needed⁸. For patients, trust is tied not just to how AI works, but to how openly and ethically it is implemented.

Where patient and physician perspectives on AI converge and diverge

When viewed together, the physician and patient perspectives reveal both points of alignment and divergence.

AI cannot succeed in healthcare without the trust of both physicians and patients.

Patients and physicians both agree that AI in healthcare:

• Has the potential to reduce administrative burden and make the healthcare experience more efficient.
• Needs safeguards against bias, improper use, and overreliance on technology.
• Needs greater oversight and governance.

Physicians and patients diverge on opinions about AI use cases: 

• Physicians are quicker to adopt AI for operational and documentation purposes, while patients remain more cautious even about administrative uses.
• Physicians worry most about the erosion of human interaction, while patients worry about loss of control, bias, and transparency.
• Patients are far less willing to accept AI in diagnostic or treatment contexts, whereas physicians, though wary, are more open to gradual clinical integration.

What This Means for the Future of AI in Healthcare

The findings underscore a crucial point: AI cannot succeed in healthcare without the trust of both physicians and patients. For physicians, that means ensuring AI tools become intuitive parts of clinical workflows and reduce — not add — to administrative burden. For patients, it means building transparency, accountability, and fairness into every use case.

Healthcare leaders and policymakers have work to do. Regulations will need to evolve in step with technology to ensure accuracy, privacy and data security. Education for both clinicians and patients will be critical to dispelling misconceptions and highlighting practical benefits. And technology companies must design AI solutions that cure complexity, not create more of it.

As the research makes clear, adoption of AI healthcare tools is not the same as acceptance. The real opportunity lies in aligning AI with the values of the people it is meant to serve — restoring time to clinicians, ensuring access to care for patients, and building a healthcare system that is smarter, fairer, and more sustainable.

Monitoring trends in AI in healthcare usage and perceptions

The story of AI in healthcare is not just about algorithms, but about trust. Patients and physicians don’t need to think exactly alike about AI. But they do need a shared sense of confidence that it is being used responsibly, transparently, and with their best interests at heart.

The athenahealth + USofC research offers a roadmap for how to get there: elevate both voices, address concerns directly, and design AI for real-world impact. If we succeed, AI won’t replace the human touch in healthcare — it will make it stronger.

Learn more about how we’re approaching AI at athenahealth.