Utah’s historic launch of the first AI doctor authorized to prescribe medication has triggered fierce backlash from the state’s medical board.
In a letter sent to the Utah Department of Commerce, the Utah Medical Licensing Board called for immediate suspension of the pilot program. The board requested further investigation into the safety and implementation of the pilot program.
Doctronic and the Utah Office of Artificial Intelligence Policy claimed the program would operate within strict parameters and with physician oversight. However, Utah’s medical board shared that it was not consulted before the state launched the pilot program to renew prescriptions for certain chronic and psychiatric conditions. The group now warns that the initiative could place patients at risk without stronger physician oversight.
“Proceeding with this agreement without consulting the Medical Board potentially places Utah citizens at risk and remains a major concern of the board,” the medical board wrote in the letter.
Medical Board Contest A First-Of-Its-Kind Regulatory Agreement
Utah officials have framed the initiative as a response to the state’s physician shortage and growing administrative burden on clinicians. In the pilot announcement, regulators said the program could help patients avoid lapses in treatment while reducing time spent on routine prescription renewals.
A commentary published in JAMA echoed the structural pressures driving the experiment. Authors Dr. Aaron Daniel and Christopher Robertson pointed to physician shortages, rising administrative workloads and barriers patients face when attempting to renew chronic medications.
The Utah Medical Licensing Board argued that prescription renewals are not simply administrative tasks. In its letter, the board said refill decisions require ongoing clinical evaluation on side effects, drug interactions, dosage adjustments and treatment effectiveness.
“Each refill requires reassessment and clinical decision-making to safely adjust doses, monitor for side effects, contraindications, or new drug interactions and ensure the medication remains effective,” the board wrote.
Board members argued that physician authorizationis necessary to reduce the risk that patients refill outdated or suboptimal prescriptions. Doctronic founders and CEOs Matt Pavelle and Adam Oskowitz acknowledged the initial risks of the program. However, they argued these risks are necessary for the U.S. health system to meet demand.
“The greater risk is continuing with a system we know does not work,” they wrote in a corporate blog post.” Expanding access requires rethinking how care is delivered.”
The dispute reflects a growing divide on whether AI should remain a physician support tool or evolve to handle autonomous clinical responsibilities.
AI Doctor Pilot Rolled Out In Phases With Mandatory Human Review Safeguards
Utah’s AI doctor system is currently not operating autonomously. The regulatory mitigation agreement for the AI rollout follows a three phase plan.
If companies meet agreed upon benchmarks and performance metrics they can move to each phase. The CEOs of Doctronic remained hopeful the program would reach fully autonomous operations.
“We’ll earn the right to operate fully autonomously by proving that we should,” the company wrote. “This is how trust gets built.”
Phase One
The program is currently in phase one. During this stage, the first 250 prescription renewals processed through the system must be reviewed by a licensed medical practitioner. A clinician must authorize each decision before it moves forward.
The agreement does not require the review to be conducted by an outside third party.
Reviewers are expected to assess clinical appropriateness, medication history accuracy, drug interaction screening, symptom assessment, and the quality of the AI system’s clinical reasoning.
Phase Two
Phase two expands oversight. During this stage, the AI tool can directly submit each prescription renewal to a pharmacist. However, every AI generated renewal decision involving the next 1,000 patients undergoes retrospective review.
Pharmacists would be able to escalate AI-generated renewals to a licensed physician working for Doctronic for further review. The agreement outlines that reviews will use the same clinical standards established during phase one.
Phase Three
Once the pilot reaches phase three, the system no longer requires review of every decision. The agreement shifts from case by case oversight to ongoing monitoring.
Under this phase, reviewers must examine 5% to 10% of all processed renewals each month. The goal is to maintain quality through structured sampling.
Escalated cases will undergo quarterly analysis by state officials. Regulators will also conduct annual reviews of system performance metrics and clinical outcomes.
The Industy Closely Monitors Nation’s First AI Doctor Experiment
Utah regulators say they are closely monitoring the state’s AI doctor pilot as debate over the program intensifies.
The initiative operates within Utah’s “regulatory sandbox,” a state-run framework overseen by the Utah Office of Artificial Intelligence Policy. The sandbox allows companies to test emerging technologies under temporary regulatory flexibility while state agencies evaluate safety, performance and compliance.
State officials say the program includes multiple safeguards intended to limit patient risk. The Office of Artificial Intelligence Policy reports that it receives monthly updates on the pilot’s performance and safety outcomes.
Under the agreement, participating companies must maintain medical malpractice insurance. The program also limits the number of prescription renewals allowed through the AI system. Once patients reach that limit, they must complete an in-person or telehealth appointment with a licensed clinician before receiving additional refills.
Regulators say the system also includes escalation protocols designed to route more complex cases to human physicians. The AI is expected to flag cases when clinical complexity exceeds preset safety thresholds, when conflicting patient information is detected or when potential medication-related complications emerge.
Security Researchers Pressure-Test The AI System As Safety Concerns Grow
So far, Utah officials say no serious safety incidents have been reported to the Office of Artificial Intelligence Policy. Still, concerns about the technology continue to grow.
Security researchers at cybersecurity firm Mindgard recently published a report identifying vulnerabilities in Doctronic’s AI platform. Researchers said they were able to manipulate the system into generating unsafe and inappropriate outputs.
According to the report, testers prompted the AI doctor to spread vaccine conspiracy theories, recommend methamphetamine as a treatment for social withdrawal and provide incorrect medication dosages.
The findings increased pressure on Utah regulators as the state’s medical board push to suspend the program pending further review.
For now, the pilot remains active. However, the dispute is quickly testing how far regulators will allow AI to participate in patient care.
