Physician AI use damages patient trust, perceived competence

Physicians who use AI are perceived as less competent and trustworthy by the U.S. public, according to new research published in JAMA Network Open.

The research examined the public perception of physicians who use AI, which is expanding rapidly in the healthcare industry, with tools like machine learning, generative AI (genAI) and agentic AI supporting an ever-growing set of use cases.

For the study, researchers created a sample of 1,276 U.S. adults based on the 2021 census. The study participants were shown fictitious social media or billboard advertisements for family physicians. The participants received similar ads but were divided into four groups: one received ads that mentioned the physician used AI for administrative purposes, the second viewed ads noting the physician used AI for diagnostic purposes, the third received ads stating the physician used AI for therapeutic purposes and the last were shown ads that made no mention of AI use.

The participants were asked to rate the physicians on a five-point scale for perceived competence, trustworthiness and empathy, as well as their willingness to make an appointment with the physician.

Participants perceived physicians as significantly less competent, less trustworthy and less empathic when they were told the physicians used AI for administrative, diagnostic or therapeutic purposes compared with physicians who did not use AI.

Additionally, participants indicated a significantly lower willingness to make an appointment with the physician if any AI use was mentioned.

"In line with prior research, our results indicate that the public has certain reservations about the integration of AI in health care," the researchers concluded.

The researchers further hypothesized that patients' negative view of AI use may result from concerns that physicians will rely too much on AI and that AI could reduce patient-physician interactions.

"From the physician's perspective it thus may be important to transparently communicate the rationale for using AI and to emphasize its potential benefits for the patient," the researchers noted.

The research adds to the growing evidence around patients' perspectives on AI use in healthcare, which could impact AI adoption in the industry.

For instance, a poll published in February 2024 revealed that 44% of patients say their trust in AI depends on the use case. However, many (43%) are unsure how AI is used in healthcare, and more than half (66%) are unsure if their provider is currently using AI as part of their clinical practice.

If given the choice, patients would like to see AI being used to automate clinician workflow and reduce provider workload (44%). They also want AI to be used to provide some diagnostic support (40%).

Anuja Vaidya has covered the healthcare industry since 2012. She currently covers the virtual healthcare landscape, including telehealth, remote patient monitoring and digital therapeutics.