While health system leaders told Becker’s they’re skeptical of the Trump administration’s push for AI physicians, they are in favor of using the technology to increase access.
HHS healthcare AI adviser Amy Gleason, who also leads the Department of Government Efficiency, told The Washington Post that discussions are underway to create a regulatory pathway for independent AI physicians, “likening the change to the decades-long process that moved self-driving cars from test tracks to cities,” according to the June 4 report.
The executive branch also supports a Utah pilot project examining AI to refill prescriptions, a program using agentic AI to communicate with patients with heart failure or who have had a heart attack, and an expedited approval process for healthcare AI chatbots, the news outlet noted.
“The pressure to move swiftly and avoid regulation to promote competitive advantage in the AI race sits in stark contrast to the conservative approach in medicine to move more slowly, develop evidence prior to implementation, and retain the confidence of our patients that we have a structured and deliberate process of evaluation before new treatments or solutions are deployed,” said Lee Schwamm, MD, senior vice president and chief digital health officer of Yale New Haven (Conn.) Health.
“The best analogy may be that while fully autonomous vehicles almost certainly make our roads safer and reduce death and disability due to motor vehicle accidents, there are still many examples of these vehicles driving into submerged roadways, halting unexpectedly, or inexplicably circling the same cul-de-sac in droves. In healthcare, those inadvertent edge case errors may produce devastating health outcomes.”
Healthcare leaders are closely following the state of Utah’s first-in-the-nation AI experiment, where technology from startup Doctronic is autonomously refilling prescriptions for nearly 200 commonly used medications. Human physicians directly reviewed the first 250 refill requests in real time and are retroactively evaluating those after that.
“We’ve been protocolizing prescription refills for a long time, and some are safe to protocolize, and some are not,” said Margaret Lozovatsky, MD, chief digital information officer of Dayton, Ohio-based Premier Health. “As much as they’re taking some of those protocols that have been developed over time and automating them, that’s a great example of taking work that has already been done within clinical decision-making that is appropriate and automating some of the busy work that happens on the back end. Which is not an AI physician — it is a tool that we can use to help decrease administrative burden.”
Dr. Lozovatsky bristles at the suggestion that there could be independent AI physicians — even in the distant future. The “heart of medicine,” as she called it, is the human relationship between a patient and a physician.
“I’m a pediatrician, I’ve been taking care of sick kids in the hospital, I can walk into a room, look at a child, and tell you if they’re sick, and that takes years and years of seeing patients and having that experience that these tools are just not able to do yet,” she said. “And I don’t know they’ll ever be at the level that is equivalent to the experience of a clinician.”
Muhammad Siddiqui, CIO of Richmond, Ind.-based Reid Health, said he finds conversational AI for cardiovascular care the most realistic idea in the short term after health systems have found success integrating ambient AI into the clinical workflow.
According to the Post, when patients contact a medical office with heart attack symptoms, a chatbot might field the call. The Advanced Research Projects Agency for Health, which is running the initiative, stated that callers will be “made fully aware that they are talking to an AI chatbot and that the program is not a substitute for a doctor or physician.” Some health systems already use an AI chatbot to perform intake and generate diagnostic and treatment recommendations (before offering to connect patients to a human physician) as part of 24/7 virtual care programs.
Darrell Bodnar, CIO of Whitefield, N.H.-based North Country Healthcare, said that while nothing would surprise him anymore about the speed of AI adoption, he would expect plenty of pushback — from health systems, industry groups and individual providers — before anything akin to an independent AI physician hits the market.
“When an AI platform recently gave me incorrect data about a car that I own, that I absolutely knew was wrong, it was inconvenient,” noted James Wellman, vice president and CIO of Gloversville, N.Y.-based Nathan Littauer Hospital & Nursing Home. “When it does that with medical data, then it could be tragic because I am not clinically trained — so how do I know it can be trusted? We want to use AI to help providers but not replace them.”
Curtis Cole, MD, chief global information officer of Ithaca, N.Y.-based Cornell University, said calling these tools AI physicians is “absurd” but that he likes the idea of moving toward a licensing-based model, with low-risk refills a good starting point.
“One of the biggest changes is liability,” Dr. Cole said. “Now physicians carry almost all liability. If a doctor uses AI and it makes a mistake, it is the doctor’s fault. If the bot gets a license and makes a mistake, the patient could sue the vendor for malpractice. That’s a game-changer … potentially.”
The Trump administration has made it clear it is taking a more hands-off approach to AI regulation than its predecessor and many states, creating a conundrum of sorts for health systems.
“That means greater flexibility to pilot and scale AI tools — especially in lower-risk areas — but also more responsibility to define our own standards for safety, validation and oversight,” said Heather Gleason, CIO of Flagstaff-based Northern Arizona Healthcare. “At the same time, if that burden increasingly falls on health systems, it will naturally slow adoption. Many of our physicians remain skeptical — particularly around liability and whether the underlying data can be trusted — which makes it even more important that we move deliberately, prove value, and build confidence as we scale.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
