Clinicians Who Tout Use of AI Risk Undermining Patient Trust, Study Finds

Physicians who advertise their use of artificial intelligence tools receive significantly lower ratings for competence, trustworthiness, and empathy from potential patients, according to a new study published in JAMA Network Open.¹

Wilfried Kunde, PhD

Courtesy of University of Wuerzburg, Germany

The research also found that patients showed significantly decreased willingness to schedule appointments with clinicians who promoted using AI for diagnostic, therapeutic, and even administrative purposes compared to those who made no mention of AI use.

Investigators from the University of Cambridge and the University of Wuerzburg, in Germany, led by Wilfried Kunde, PhD, professor of psychology at the latter, underscore the rapid traction AI has gained in science and medicine, but cite research suggesting that public skepticism in AI integration into health care is growing.² To date, they point out, research has examined attitudes toward medical AI tools and AI-generated advice,³ but there is very little known about how the public perceives physicians themselves who use these technologies.

Wilfried et al conducted an online survey in January 2025, recruiting 1,276 US adults using quota sampling based on 2021 census data. Participants included 680 women (53.3%), 584 men (45.8%) and had a mean age of 46.2 years. The study used a randomized design where participants viewed fictitious advertisements for a family doctor that might be seen on billboards or on social media.¹

After dividing participants into 4 groups, investigators showed each group an identical advertisement with one specific difference: whether the physician mentioned using AI for administrative tasks, diagnostic purposes, therapeutic applications, or made no AI reference at all. Participants rated each physician on competence, trustworthiness, empathy, and willingness to make an appointment using 5-point scales.¹

FINDINGS

Compared with the control group, which saw the ad with no mention of AI, the groups who saw ads referencing AI use rated the physician lower in all 4 domains.¹

Competence

Control group 3.85 points vs:

• Administrative AI (3.71 points, P =.04)
• Diagnostic AI (3.66 points, P =.007)
• Therapeutic AI (3.58 points, P <.001)

Trustworthiness ratings showed even more pronounced differences

Control group 3.88 points vs:

• Administrative AI (3.66 points, P =.001)
• Diagnostic AI (3.62 points, P <.001)
• Therapeutic AI (3.61 points, P <.001)

Empathy ratings followed similar patterns

Control group (4.00 points) vs:

• Administrative AI (3.80 points, P =.001)
• Diagnostic AI (3.82 points, P =.005)
• Therapeutic AI (3.72 points, P <.001).

Willingness to schedule an appointment with a physician was the condition affected most substantially by the mention of use of AI

Control group (3.61 points) vs:

• Administrative AI (3.32 points, P <.001)
• Diagnostic AI (3.16 points, P <.001)
• Therapeutic AI (3.15 points, P <.001)

According to the study, there were no significant differences observed between the AI conditions for any rating dimension.

“In line with prior research, our results indicate that the public has certain reservations about the integration of AI in health care,” the authors wrote. “While the present effect sizes are relatively small, in particular regarding AI use for administrative purposes, they may be highly relevant as trust in health care practitioners is closely linked to subjective treatment outcomes.”¹

In their discussion of the study's limitations Wilfried and colleagues acknowledge the use of hypothetical scenarios, the artificial nature of stimuli, and potential selection bias from recruiting participants willing to participate in such experiments. Those deficits, however, lay groundwork for future research that would benefit from examining more realistic settings, they wrote, and exploring potential confounders, eg, patients' prior experience with AI and digital tools.¹

Although reasons for the skepticism telegraphed in the study are unclear, the researchers suggest that potential patients may be increasingly concerned that clinicians are coming to rely too heavily on AI and moving farther away from the high-touch ethos that has traditionally characterized "good" medical care.

As an antidote to any underlying or nascent patient distrust, the research team said it is important that clinicians "transparently communicate the rationale for using AI and emphasize potential benefits for the patient."¹

References

1. Reis M, Reis F, Kunde W. Public perception of physicians who use artificial intelligence. JAMA Netw Open. Published online July 17, 2025. doi:10.1001/jamanetworkopen.2025.21643

2. Fanous A, Steffner K, Daneshjou R. Patient attitudes toward the AI doctor. Nat Med. 2024;30(11):3057-3058. doi:10.1038/s41591-024-03272-4

3. Longoni C, Bonezzi AK, Morewedge CK. Resistance to medical artificial intelligence. J Consum Res. 2019;46(4):629-650. doi:10.1093/jcr/ucz013