AI and doctors: Genuine excitement and lingering concerns

Doctors are expressing some excitement about the use of AI, but some of them still have concerns ranging from patient privacy to legal liability before using it more widely.

A new American Medical Association survey released Thursday outlines some of the mixed emotions physicians have with artificial intelligence, or augmented intelligence as outlined in the group’s report.

Nearly two-thirds of the physicians surveyed (65%) said they see at least some advantage in using AI in their practices.

So far, most doctors aren’t using AI, the survey found. Only 38% of the physicians surveyed said they were using AI in their practices. Most of those using AI said they were using tools to assist with documentation and translation, along with some assistance in diagnosis.

Doctors were asked if they were more excited about the potential of AI, or if their concerns outweighed their enthusiasm. In a telling response, 41% said they were equally excited and concerned about using AI. In addition, 30% said they were more excited than concerned, while 29% were more concerned than excited.

Nearly three out of four (72%) surveyed said they were most excited about the potential of AI to improve their diagnostic capabilities, while many (69%) cited the potential gains in workplace efficiency. A solid majority (61%) said they saw potential for better clinical outcomes.

Jesse Ehrenfeld, president of the AMA, noted the enthusiasm for AI, particularly in dealing with what he called “the crushing administrative burdens that plague modern medicine.”

“Physicians are optimistic about the advantages that properly designed AI-enabled tools can have for patient care, and nearly two-thirds of physicians see an advantage to AI if key requirements are met,” Ehrenfeld said in a statement.

However, 41% of doctors said they were concerned about patient privacy, while 39% said they thought AI would have a negative impact on the doctor-patient relationship. They emerged as the leading concerns in the survey.

Doctors said they had the most enthusiasm for using AI in documentation of encounters with patients and medical charts, and automating insurance pre-authorization.

In terms of what it will take to get doctors to use AI more widely, physicians expressed the most concern in two areas: assurances of data privacy, and assurances that physicians won’t be held liable for errors of AI models. The survey found 87% of doctors labeling both areas as important concerns. Nearly as many doctors (86%) pointed to AI being covered by malpractice insurance.

The AMA surveyed 1,081 doctors in August, including a mix of specialists and primary care physicians.

Healthcare leaders have acknowledged AI’s potential to transform healthcare, particularly in being able to identify patients at risk of serious illness. AI has been a hot topic at healthcare conferences over the past year.

However, critics point to the need for strong guardrails regarding the use of AI. They noted that some AI chatbots have provided answers to healthcare questions indicating racial bias. The Coalition to End Racism in Clinical Algorithms has been working with New York health systems to improve racial equity in algorithms.

President Biden’s administration issued an executive order on AI in healthcare in November. The Biden administration said it would offer grants to expand the use of AI, but also said the government will develop a safety program to deal with unsafe healthcare practices involving the use of AI. The order also calls for the developers of AI systems to share their safety test results with the government.