Utah has approved three new pilots for its healthcare AI program, expanding experiments that could eventually allow the technology to issue prescriptions and care plans without direct clinician review. The state’s Office of Artificial Intelligence Policy gave final approval on Friday, even as debate over the approach continues.
The pilots include a partnership with Nolla Health, which will test whether its dermatology app, using machine-learning analysis of patient photos, can prescribe and refill topical medications for mild-to-moderate acne. In the early stage, two physicians will review every prescription before it is approved, though later phases may reduce that oversight. The program is aimed at adults living in Utah who have the condition.
Two other pilots were also approved: August AI will use its system to renew prescriptions for patients with chronic conditions such as diabetes or hypertension, while Expect Fitness will deliver virtual pelvic-floor therapy via video sessions for women experiencing urinary incontinence, pelvic pain, or postpartum issues. All three require written approval from regulators before moving to phases with less human involvement.
None have launched yet, though Utah’s AI Learning Laboratory, created by state law in 2024, provides a framework for testing such tools under strict safeguards. Companies must sign agreements that spell out boundaries, consumer protections, and reporting requirements before they can participate. The lab seeks to balance innovation with oversight, letting private firms experiment while mitigating potential risks.
Criticism from doctors and other stakeholders persists. Last year, Utah’s medical licensing board asked for a suspension of a Doctronic pilot that allowed AI to renew medications without clinician oversight, citing a lack of transparency and concerns about outdated prescriptions. Regulators declined to pause the program, arguing that clinicians remained involved in the early stages.
Experts have highlighted broader risks. A commentary in JAMA Health Forum noted that Doctronic’s contract does not clearly define liability for AI-related prescribing errors. Despite these concerns, the regulators continue to expand AI trials. Last month, the Office of Artificial Intelligence Policy also signed master agreements with University of Utah Health and Intermountain Health, setting terms for future collaborations.
While these trials focus on specific conditions, the long-term implications extend beyond dermatology and chronic-care medication renewal. The state’s approach reflects a growing trend in healthcare AI, testing automation in controlled settings before wider adoption. For now, oversight remains a key condition, and the line between AI-driven decisions and human intervention is gradually shifting.
