🔎︎
Utah is testing the use of autonomous artificial intelligence for prescription renewals, according to a June 18, 2026 report by Anna Wilde Mathews in The Wall Street Journal. The project, launched in January with technology from the startup Doctronic, is designed to eventually allow an AI product to renew certain prescriptions without prior physician review. The task is typically performed by physicians. Utah adults can request refills for a limited list of medications, including cholesterol drugs and antidepressants such as Lipitor and Prozac, with the stated goal of improving access to medication.
The project operates through a regulatory sandbox overseen by the Utah Department of Commerce, and state officials describe it as a conservative, phased program. In the current phase, every AI-generated prescription is reviewed by a physician before it is filled. According to results released by the state, the AI system approved 72% of renewal requests and escalated the remaining requests to a doctor for reasons such as the need for lab work. Among AI-approved renewals, reviewing physicians concurred 91% of the time, and 3% of AI approvals were ultimately rejected by two physician reviewers. The first phase is set to continue until at least 250 prescriptions are accumulated for each drug class.
Some physicians have raised objections. Most members of the state medical licensing board signed a letter seeking suspension of the project on safety grounds, citing what they described as inadequate vetting and the risk of continuing medications that may cause side effects or no longer be appropriate. Board members also raised questions about liability and said the technology lacks the transparency, training, and licensure requirements applied to human clinicians. The American Medical Association has broadly opposed using AI as a replacement for physicians, with its chief executive warning that such use could lead to AI systems making more complex medical decisions.
Doctronic said its system follows physician-created clinical guidelines, that its reviewing doctors exercise independent judgment, and that it will publish peer-reviewed pilot results. State officials said they required malpractice insurance for the AI system and reported no major safety incidents so far. The article notes that states are taking different approaches to regulating autonomous clinical AI, with some restricting it and others moving toward licensing it.
