Utah Is Now Letting an AI System Prescribe Psychiatric Drugs Without a Doctor

Utah has become only the second U.S. state to formally authorize an AI system to prescribe psychiatric medications without physician oversight. The state's authorization covers prescription and refill decisions for a defined set of psychiatric drugs, allowing the AI to function in a clinical capacity that has historically required a licensed physician's judgment.

State officials frame the decision around access: Utah faces significant shortages of psychiatrists, particularly in rural areas, and the authorization is intended to extend the reach of psychiatric care to patients who would otherwise face months-long wait times or no access at all. On cost grounds, AI-assisted prescribing can reduce per-appointment overhead substantially.

What Physicians Are Warning

The medical community's concerns are specific, not reflexively anti-technology. Psychiatric medication decisions are complex in ways that scale poorly to algorithmic treatment: drug interactions are numerous and sometimes idiosyncratic, patient response curves vary widely, and the diagnostic signals that inform medication adjustments — affect, body language, the texture of a patient's account of their symptoms — are difficult to assess through a digital interface. Physicians who have reviewed Utah's authorization describe the AI system as "opaque" — they cannot audit the model's reasoning or identify why it made a specific prescribing decision.

The opacity concern is not minor. In medicine, the ability to understand why a treatment decision was made is foundational to accountability, error correction, and informed consent. A prescribing system that cannot explain its reasoning is not just a regulatory problem — it is a patient safety problem.

The Kintsugi Parallel

The same week Utah's authorization became public, California-based startup Kintsugi — which spent seven years developing AI designed to detect signs of depression and anxiety from speech — announced it is shutting down after failing to secure FDA clearance in time. The contrast is stark: one jurisdiction is extending prescribing authority to AI systems, while the FDA's clearance process forced a carefully-developed diagnostic tool out of the market. The regulatory landscape for AI in clinical care is not coherent — it is a patchwork of state-level experiments operating in a federal vacuum.

Where This Goes

The question is not whether AI will play a larger role in psychiatric care — the shortage math makes that outcome likely regardless of the policy debate. The question is whether the governance structures being built now will be adequate to catch the failures that inevitably occur. Utah is running that experiment in real time, on real patients, with an opaque system. The results will matter for every state watching to see what happens next.