Researchers explore legal and ethical questions raised by Utah AI prescription pilot
A New England Journal of Medicine article explores legal, ethical and patient safety questions raised by a Utah program testing artificial intelligence to independently renew prescriptions for certain chronic conditions.
Ravi B. Parikh, MD, MPP
A first-in-the-nation program in Utah is testing whether artificial intelligence can independently renew prescriptions for certain chronic conditions, and a new article in New England Journal of Medicine explores the legal, ethical and patient safety questions the initiative raises.
The commentary includes Ravi B. Parikh, MD, MPP, associate professor in the Department of Hematology and Medical Oncology at Emory University School of Medicine and a researcher at Winship Cancer Institute of Emory University, along with co-authors Sara Gerke, associate professor of law and Richard W. & Marie L. Corman Scholar at University of Illinois College of Law, and I. Glenn Cohen, deputy dean and professor at Harvard Law School.
Utah’s 12-month pilot launched in January 2026 through a state “regulatory sandbox,” a framework designed to test emerging technologies under limited conditions. The program uses an AI system developed by Doctronic and applies to nearly 200 commonly used medications, including treatments for high blood pressure, diabetes and depression. After an initial physician review period of 250 cases, the system is expected to begin making prescription renewal decisions without direct human oversight.
The authors note that AI-powered prescription renewal may offer benefits in carefully defined settings, particularly if it helps improve medication adherence for people managing chronic disease. At the same time, they write that the Utah program highlights the need for careful oversight as more autonomous tools enter health care.
Among the issues raised in the article are potential risks when AI is used for medications that may require frequent dose adjustments or for patients whose medical status can change rapidly. The authors also examine whether the system should undergo FDA premarket review as a medical device and whether AI-based prescribing could create concerns under federal misbranding laws, which require prescribing to be carried out by a licensed practitioner.
The article also explores the evolving relationship between state innovation efforts and federal regulation, a topic likely to grow in importance as AI becomes more integrated into patient care.
“Such systems hold a lot of potential benefit for patients,” Cohen says. “But especially as the first-in-the-nation, it is important for patients that the developers consider all the legal and ethical issues raised.”
“Artificial intelligence may help streamline some routine aspects of care, but health care decisions still require careful attention to patient context, safety and accountability,” Parikh says. “As these technologies evolve, it is important to evaluate both their promise and their risks.”
The article adds to a growing national discussion about how AI can support clinicians and patients while maintaining appropriate safeguards, transparency and trust.