Outrage of the Month: Artificial Intelligence and Drug Prescribing
Health Letter, July 2026
By Robert Steinbrook, M.D.
Director, Public Citizen's Health Research Group
(A version of this column was published in the July 2026 issue of Worst Pills, Best Pills News)
In January 2026, Utah launched a pilot program with an artificial intelligence (AI) company to allow the firm’s proprietary AI healthcare platform to autonomously refill prescription medications for people in the state with certain common conditions, such as diabetes and high blood pressure. The partnership is with Doctronic, which describes itself as “the world’s #1 AI doctor” and boasts that its AI chatbot is trained “exclusively on peer-reviewed medical sources, so the info we provide is medical-grade and doctor-approved.” About 190 medications are eligible for renewal.
AI is a software application, not a licensed physician or other medical professional. Software should not be refilling prescriptions autonomously, nor should it be identifying itself as an “AI doctor” that “you can talk to me just like you would your regular doctor.” The term “AI doctor” misleads consumers and debases the meaning of the word “doctor.”
Why is this pilot program legal? In 2024, Utah created an “AI Learning Lab” overseen by the state’s Office of Artificial Intelligence Policy within the Department of Commerce and the Division of Professional Licensing. To test new programs, the AI learning lab “can temporarily waive certain professional licensure and scope-of-practice laws,” including requirements for prescriptions to be renewed by a human clinician. Without this waiver, any drug dispensed without a prescription from a licensed practitioner would have been considered misbranded, making its sale a crime.
There are legal uncertainties, however. It remains unclear “whether the word ‘practitioner’ can apply to a nonhuman entity,” such as autonomous AI, an uncertainty that Congress or courts may eventually need to resolve.
The program is also not overseen by the Food and Drug Administration (FDA). Although states, not the federal government, regulate the practice of medicine, the FDA regulates medical devices. According to a recent article in the New England Journal of Medicine, the AI drug prescribing chatbot “should probably be considered a device under the Federal Food, Drug and Cosmetic Act.” The article also noted that, “If the FDA doesn’t initiate action against Doctronic, however, nothing will happen. The FDA may decide to look the other way and allow this experiment to continue, leaving clinicians and patients without safety assurances amid an extraordinary expansion of autonomy for AI.”
The Utah Medical Licensing Board, which was not consulted before the program was launched, has called for the suspension of the state’s AI-enabled prescription renewal systems pending further study. Public Citizen agrees. We have urged Utah to suspend its AI-enabled “prescription renewal systems pending further review of their legality, safety, accountability, and compliance with existing standards governing the practice of medicine.”
As of late June 2026, the Utah pilot program is continuing. Although putting the genie back in the bottle would be a daunting task, deploying AI to autonomously perform prescribing and other core medical functions without robust evidence of safety and effectiveness sets a dangerous precedent. Adequate safeguards and accountability measures must be in place. State medical boards, not commerce departments, should decide when, how and for what purposes AI can be deployed to facilitate clinical practice and improve patient care.