What's Happening?
Utah has launched a pilot program allowing artificial intelligence (AI) to issue prescriptions for mild to moderate acne, marking the first time in the U.S. that AI could autonomously prescribe medication.
The program, run by Nolla Health under a 12-month agreement with the state’s Office of Artificial Intelligence Policy, involves adults completing a questionnaire and face scan, after which the AI system selects from pre-approved topical treatment plans. Currently, physicians review and approve every prescription, but the program aims for autonomous AI prescribing in later stages, with physician oversight reduced to reviewing a sample of prescriptions. This development has drawn criticism from the American Medical Association (AMA), which asserts that prescribing medication is an essential part of medical practice and that AI should support, not replace, physicians. Concerns include the lack of randomized controlled trials to prove efficacy and safety, and the potential for AI to miss nuances in skin conditions, especially in people of color.
Why It's Important?
This pilot program in Utah represents a significant step towards integrating AI into direct patient care, particularly in the realm of prescription writing. While proponents argue it could address gaps in access to dermatological care, especially in underserved areas, the initiative raises fundamental questions about patient safety, medical liability, and the definition of medical practice. The AMA's strong stance against autonomous AI prescribing highlights a broader debate within the medical community regarding the appropriate role of AI. If AI is allowed to prescribe without direct physician oversight, it could set a precedent for other medical conditions, potentially transforming healthcare delivery but also introducing new risks. The lack of randomized trials and concerns about AI's ability to account for diverse patient needs underscore the importance of rigorous testing and ethical considerations before widespread adoption.
What's Next?
The Utah pilot program is structured in phases, with the ultimate goal of autonomous AI prescribing. Currently, two physicians review each prescription, but future phases will reduce direct oversight, eventually leading to physicians reviewing only a sample of prescriptions. Advancement to each subsequent phase requires at least 95% agreement between the AI and physicians, no serious adverse events, and state approval. The state's Office of AI Policy states that independent evaluators will verify developers' claims. However, the AMA is advocating for transparent, auditable evidence of safety and efficacy before AI tools are used for prescribing. The outcome of this pilot will likely influence future regulatory decisions and the broader adoption of AI in prescription management across the U.S., potentially leading to increased scrutiny and calls for more robust clinical trials.
Beyond the Headlines
The Utah AI prescription pilot program delves into profound ethical and legal implications. The core debate revolves around whether AI can truly replicate the nuanced judgment of a human physician, especially when considering diverse patient demographics and potential side effects. The concern about AI's ability to diagnose and treat skin conditions in people of color, where presentations can differ, highlights the critical issue of algorithmic bias and its potential to exacerbate health disparities. Furthermore, the question of who is accountable in cases of AI-induced adverse events—the AI developer, the prescribing physician, or the state—remains largely unanswered. This pilot could trigger a re-evaluation of medical licensure, malpractice laws, and the very nature of the doctor-patient relationship in an increasingly AI-driven healthcare landscape, pushing for new regulatory frameworks to ensure patient safety and equitable care.








