Artificial intelligence is advancing across medicine at a remarkable pace. New tools are helping physicians reduce administrative burdens, improve communication and support clinical decision-making.
But some are now performing functions that increasingly resemble the practice of medicine itself, while the guardrails needed to ensure safety, accountability and public trust are not keeping pace. That gap should concern everyone.
This is why Utah’s autonomous prescribing pilot deserves national attention. Utah has built a reputation as a state willing to experiment with new approaches to technology and regulation. That willingness to innovate can create important opportunities to improve access and efficiency in healthcare. Like many states, Utah is also confronting real challenges around access to care and medication adherence, problems that can negatively affect patient outcomes and deserve thoughtful solutions.
But when innovation begins to intersect with the practice of medicine, the stakes are different.
In medicine, trust is not a feature. It is the foundation. Patients must trust that the technologies influencing their care are safe, effective and subject to meaningful oversight. Physicians must trust that these systems have been rigorously evaluated before they are integrated into patient care.
The Utah pilot allows an AI system to autonomously renew prescriptions under certain circumstances. While prescription renewals may seem routine, medicine is rarely as straightforward as it appears.
Patients do not always fit neatly into algorithms, protocols or predefined rules. A medication refill may involve a new symptom, an emerging side effect, a drug interaction or a change in laboratory results or other circumstances that alter what is clinically appropriate. This is where clinical judgment matters.
AI systems can be effective at operating within defined parameters. What they cannot do is fully replicate the judgment that comes from understanding the complete context of a patient’s health, history and needs. That is why meaningful physician oversight must remain central to the use of AI in patient care.
The question is not whether AI belongs in healthcare. The question is whether we will put the right safeguards in place before these technologies become deeply embedded in patient care.
As states explore new models for AI-enabled care, it is critically important for physicians and state medical boards to be part of the conversation from the beginning. If a technology is being authorized to perform functions that resemble the practice of medicine, then the entities responsible for overseeing the practice of medicine should have a seat at the table.
What is happening in Utah is not unique to this state. It is a preview of questions every state will soon face.
Should autonomous clinical AI systems undergo review by the Food and Drug Administration? Who is accountable when an AI system causes patient harm? What level of evidence should be required before these technologies are deployed at scale?
These questions demand answers.
At a minimum, AI systems performing clinical functions should undergo rigorous clinical validation. Developers should accept clear accountability for patient harm and should not be able to disclaim responsibility through contractual fine print. Strong post-deployment monitoring and transparent reporting mechanisms should be required so that patients and physicians understand both the benefits and risks of these technologies.
The American Medical Association recently strengthened its policies around AI-enabled tools to better support patient-centered care and emphasize that AI should serve as an assistive tool, not an autonomous decision-maker, in medical practice.
Calling for appropriate guardrails for AI technology is not opposition to innovation. It is exactly the opposite.
AI has tremendous potential to improve healthcare, reduce burdens on physicians, expand access, and help patients. But its success will ultimately depend on whether it earns public trust.
Trust is built through evidence, transparency and accountability. Utah has an opportunity to help lead the national conversation about how AI should be integrated into patient care.
But leadership in healthcare innovation is not measured by how quickly technology can be deployed. It is measured by whether physicians and patients trust it.
Healthcare AI should move forward, but not simply at the speed of investment. It should move at the speed of safety and, ultimately, trust.
(American Medical Association) John Whyte, MD, MPH, is chief executive officer of the American Medical Association.
John Whyte, MD, MPH, is chief executive officer of the American Medical Association.
The Salt Lake Tribune is committed to creating a space where Utahns can share ideas, perspectives and solutions that move our state forward. We rely on your insight to do this. Find out how to share your opinion here, and email us at voices@sltrib.com.