Patients Trusted Doctors Less Who Disclosed AI Use

A 2025 study found that patients rated physicians as less warm and competent when they acknowledged using AI.

Updated on Oct. 5, 2026 in Artificial Intelligence

Bold flat-color editorial illustration of a medical diagnostic lamp on a table, representing clinical trust in the age of AI. — A 2025 study of 1,030 U.S. adults found that patients are less likely to trust or recommend physicians who disclose the use of artificial intelligence tools. AI Illustration. Upload story photo >

Live Poll

Would you trust a doctor less if they disclosed that they use artificial intelligence?

In November 2025, a study of 1,030 US adults revealed that patients are less likely to recommend or pay for appointments with doctors who disclose AI usage. Participants consistently viewed these physicians as having lower warmth and competence.

Why it matters

The findings suggest that patients may interpret AI disclosure as a doctor outsourcing their clinical judgment, which directly undermines perceptions of professional expertise. Younger patients, in particular, demonstrated more skepticism regarding the integration of these tools into care.

The study utilized a factorial design across 1,030 participants, including 674 White, 123 Black, and 72 Asian respondents. Findings indicated an AI-use penalty regarding willingness to pay and travel, which remained consistent across physician age, race, and gender.

The details

Researchers presented participants with randomized advertisements for family doctors to measure how the mention of AI tools influenced patient trust. The resulting data showed that disclosing AI use led to diminished willingness to recommend the provider and a desire to travel shorter distances for care.

The study data was collected throughout November 2025.

The Big Picture

This study highlights how patient trust acts as a significant barrier to the adoption of the National Institutes of Health's clinical decision support research initiatives in primary care settings. It marks a departure from purely technical evaluations of software efficacy by centering the social psychology of the doctor-patient relationship.

Patients may find themselves choosing providers based on personal comfort levels rather than technical capabilities if they hold negative views on AI integration. This may lead to longer commute times for patients who prioritize human-only clinical diagnostic processes.

The takeaway

Healthcare providers should be mindful that transparency regarding technology usage may inadvertently lower patient confidence during initial consultations. Improving patient education on how AI assists rather than replaces clinical judgment could help mitigate these negative perceptions.