# Utah Opens AI Prescription Path > Utah authorizes first state AI prescription system with physician oversight. - Title: Utah Opens AI Prescription Path · Dallas Express | Shortify - Summary: Utah authorizes first state AI prescription system with physician oversight. Utah launched a state-regulated AI system for acne prescriptions with physician… - Keywords: utah, prescription, nolla health, acne treatment, physician review, healthcare ai, texas, news, politics, Opens, Path, Dallas Express - Source: Dallas Express — https://dallasexpress.com/health/can-ai-prescribe-medicine-utah-just-opened-the-first-state-path-for-initial-prescriptions - Author: Dallas Express - Published: 2026-10-06T17:00:59+00:00 - Read time: 2 min - Topics: utah, ai, prescription, nolla health, acne treatment, physician review, healthcare ai, texas, news, politics ## Utah AI Prescription Pilot Utah opened the nation's first state-regulated path for AI to issue initial prescriptions without immediate doctor review. > Utah Office of Artificial Intelligence Policy authorized Nolla Health's acne-treatment pilot under an agreement signed September 22. ## Enrollment Starts October 5 Utah residents age 18 and older began enrolling in Nolla Health's AI acne prescription service on October 5. > Nolla announced on October 5 that Utah residents age 18 and older could begin enrolling in the service. ## Two-Phase Physician Review The first stage requires two Utah-licensed physicians to review every AI-generated prescription before pharmacy delivery. > The first stage still requires two Utah-licensed physicians to review every AI-generated prescription before the system sends it to a pharmacy. ## Pilot Conditions The AI must treat at least 100 patients, maintain 95% agreement with physicians, and miss no safety stops within four weeks. > Under the agreement, the AI can begin submitting prescriptions directly only after the pilot runs for at least four weeks, treats at least 100 patients, maintains at least 95% agreement with physicians, misses no required safety stops and records no serious adverse events. ## Stage 2 and 3 Progression Stage 2 would let AI submit prescriptions directly with weekly physician review, while Stage 3 reduces review to monthly sampling of 10% of patients. > If the pilot advances, Stage 2 would let the AI submit prescriptions directly while physicians review every case afterward at least weekly. Stage 3, which requires at least 750 cumulative patients, would reduce routine physician review to a monthly sample of at least 10% of enrollment. ## Acne Treatment Scope The AI treats mild to moderate acne with topical products like tretinoin and spironolactone, avoiding oral drugs and severe acne treatments. > Nolla's system uses a structured medical intake, five-angle facial images and predefined clinical protocols to assess adults with mild to moderate acne. ## Hard Stops for Safety The system rejects prescriptions if pregnancy, breastfeeding, immune issues, medication reactions, or poor image quality are detected. > The system also uses hard stops. Pregnancy or plans to become pregnant, breastfeeding, a weakened immune system, a previous reaction to a listed medication, severe acne and poor-quality images can force the system to reject the automated path or route the patient to a physician. ## Safety Event Reporting Nolla must report missed drug allergies, interactions, incorrect dosing, and inadequate diagnostic workups as safety events. > Nolla's agreement treats a missed drug allergy, contraindication, drug-drug interaction, incorrect dose or frequency, or failure to recommend appropriate diagnostic workup or physician follow-up as a reportable safety event. ## Medication Error Statistics The CDC estimates 1.5 million people visit emergency departments annually due to medication errors, leading to nearly 500,000 hospitalizations. > The Centers for Disease Control and Prevention estimates that adverse drug events send more than 1.5 million people to U.S. emergency departments each year and lead to nearly 500,000 hospitalizations. ## 2005 Prescription Study A 2005 study found 7.6% of prescriptions contained errors, with 3% having potential for patient injury. > A widely cited outpatient study found prescribing errors in 7.6% of 1,879 prescriptions reviewed at four Boston primary-care practices. ## Physician Resistance The Utah Medical Licensing Board urged suspension of an earlier AI prescription-renewal pilot with Doctronic in April. > In April, the Utah Medical Licensing Board urged the state to suspend an earlier AI prescription-renewal pilot with Doctronic. ## AMA Cautious Stance The American Medical Association states AI cannot replace physician judgment in healthcare decisions. > The American Medical Association took a similarly cautious position in June. The organization adopted policy calling for AI to support rather than replace physician judgment. ## Nolla's 96% Agreement Nolla reported 96% agreement with licensed clinicians on AI treatment recommendations in the last 1,000 treatments. > Nolla points to encouraging early data in its Utah proposal. The company said licensed clinicians agreed with its AI-generated treatment recommendations in about 96% of the last 1,000 treatments. ## Texas Regulatory Sandbox Texas created a 36-month AI testing program for healthcare with physician review requirements. > Texas has already built a legal structure that could support a similar experiment. House Bill 149, which took effect January 1, created an Artificial Intelligence Regulatory Sandbox Program that expressly includes health care among the sectors eligible for controlled testing. ## Key takeaway Utah launched a state-regulated AI system for acne prescriptions with physician review, requiring 95% agreement with doctors before full autonomy.