Utah Greenlights AI Systems to Prescribe Medications Without Real-Time Doctor Oversight

Utah's Office of Artificial Intelligence Policy has approved three healthcare AI pilots designed to autonomously prescribe and refill medications without real-time clinician review, marking a significant expansion of the state's controversial AI healthcare sandbox. The approvals include a dermatology app for acne treatment, a chronic condition medication renewal system, and a virtual pelvic floor therapy platform—all of which will eventually operate with minimal human supervision.

Autonomous Prescribing Moves Forward Despite Professional Concerns

The newly approved pilots represent Utah's continued push into AI-driven healthcare automation. Nolla Health, a dermatology application, will prescribe initial medications and refills from a limited list of topical drugs for adults with mild-to-moderate acne in Utah. Initially, two physicians will review every prescription before it's sent, but later stages will involve only retrospective review—meaning prescriptions will be issued before a doctor examines them.

August AI will focus on renewing prescriptions for patients with chronic conditions, while Expect Fitness will provide virtual pelvic floor therapy for women addressing urinary incontinence, pelvic pain, and postpartum pelvic floor problems. None of the newly approved pilots have launched yet, and all require written regulatory approval before advancing to phases with reduced human oversight.

These approvals come despite ongoing controversy surrounding earlier pilots. The Doctronic pilot, launched earlier in 2024 to renew medications without clinician oversight, faced significant opposition from Utah's medical licensing board. Michelle Mello, a professor of law and health policy at Stanford University, raised concerns in a JAMA Health Forum commentary about unclear liability for AI prescribing errors in the Doctronic contract.

Utah's AI Sandbox Creates New Regulatory Pathway

Utah established its AI Learning Laboratory—commonly called an AI sandbox—through state law in 2024 to provide a supervised testing environment for AI healthcare tools. The regulatory structure allows private companies to sign agreements with the state defining boundaries, safeguards, and reporting requirements for their pilots.

Major healthcare systems have already committed to the initiative. University of Utah Health and Intermountain Health signed master agreements in January 2025, signaling institutional support for the experimental framework despite physician concerns about clinical oversight, liability, and patient safety when AI systems operate with minimal human supervision.

The phased approval process requires companies to demonstrate safety and effectiveness at each stage before achieving full autonomy. However, critics argue that the framework diminishes physician and medical board influence over AI oversight policies, potentially prioritizing innovation over patient protection.

Implications for Healthcare's AI Future

Utah's regulatory sandbox creates a precedent that may influence other states considering similar frameworks for autonomous AI prescribing. For healthcare providers, the pilots could reduce workload for routine prescriptions but raise unresolved questions about liability and clinical judgment in medication management. Health tech companies gain a unique pathway for real-world validation and market entry, particularly for prescription management and specialty telehealth applications.

For Utah residents, the pilots promise improved convenience and access to specialized care in dermatology, chronic disease management, and women's pelvic health. However, patients face uncertain liability protections and potential risks from AI errors in medication management without real-time physician oversight—concerns that remain unaddressed as the state moves forward with autonomous prescribing technology.

Sources

1. Utah approves slew of health AI pilots for drug management, women's health — Healthcare Dive