Utah residents can now skip the doctor's office and get their prescriptions renewed by an artificial intelligence chatbot, a first-in-the-nation experiment that the state's own medical licensing board never approved and has formally asked to shut down.
The program, operated by a health-tech startup called Doctronic, launched quietly in January under a state "regulatory sandbox" that lets officials waive existing laws for AI companies offering what they consider promising technology. Patients visit a Doctronic website, answer questions from the chatbot about their prescriptions and medical history, and walk away with a refill sent to their local pharmacy. No physician visit required.
For a country already wrestling with how fast to let artificial intelligence into high-stakes decisions, the Utah AI prescription refill program is a live test case, and the early returns are raising alarms among doctors, lawyers, and the federal government's own drug regulators.
Dr. Alan Smith, a family physician who heads Utah's state medical licensing board, told the Associated Press that he and his colleagues learned the program existed only when its January launch was covered by reporters:
"We were essentially told: 'Yes this is going on. And no, you don't have a say in it.'"
By March, eleven members of the board had signed a letter to the state calling for the program to be halted. Their concern: automatically renewing medicines that can carry side effects or dangerous drug interactions puts patients at risk.
Smith pointed to Doctronic's list of 190 medications eligible for AI-managed refills, a list that includes blood thinners. He warned that a prescription that made sense six months ago may not make sense today.
"Many times when I see people after six months I find that their medical history or situation has changed. Just because something was prescribed before does not mean it's appropriate now."
The state has since removed several medications from the eligible list in response to safety concerns, including a drug used to treat irregular heartbeats. But the program itself continues.
Doctronic's operations are overseen by a five-member board of AI specialists. None of them are doctors. During the initial phase, human physicians still review all refill orders, but the company expects to transition to fully automated refills soon, with no doctor in the loop at all.
The only published performance data comes from Doctronic itself: a company-authored study, not independently reviewed, that examined 500 telehealth consultation records. Doctronic's diagnoses matched those of human doctors 80 percent of the time. The company says it plans to publish peer-reviewed studies later this year.
An 80 percent match rate might sound impressive in a Silicon Valley pitch deck. In medicine, it means one in five diagnoses didn't match. For a system preparing to operate without physician oversight, that gap matters.
Dr. Eric Bressman of the University of Pennsylvania framed the situation bluntly, as reported by Breitbart:
"We have crossed a threshold in terms of giving something that is not human a medical license, whether or not we want to call it that."
Bressman also noted that Utah's regulatory sandbox board appeared to accept Doctronic's claims largely on trust. "Mostly they're accepting the company's word on good faith that they're up to the task," he said.
State and federal laws limiting prescribing authority to licensed medical professionals have underwritten American medicine for over a century. Utah's regulatory sandbox effectively suspends those rules for companies the state deems innovative enough to warrant an exception.
Doctronic co-founder Dr. Adam Oskowitz described the company's vision in expansive terms, saying its goal is to "meet patients where they need healthcare." He envisions a future where routine medical tasks are offloaded to AI, allowing doctors to manage thousands more patients. He added: "We try not to get too deep into the weeds on the regulatory side."
That last line deserves a second read. A company operating in one of the most heavily regulated sectors in American life, one where a wrong dosage can kill, says it tries not to get too deep into the regulatory weeds. The medical licensing board that exists specifically to protect patients in that state was not consulted before the program went live.
When government programs expand into sensitive areas of healthcare, whether it's Medicare's new GLP-1 weight loss drug initiative or an AI chatbot writing prescriptions, the questions about oversight, cost, and patient safety follow close behind.
The Food and Drug Administration, the federal agency with the clearest authority to intervene, has so far declined to do so. An FDA spokesperson said the agency "is committed to encouraging medical innovation and helping bring promising new technologies to patients, while keeping safety at the center of every decision." But the spokesperson also confirmed the FDA has not authorized any AI chatbots.
The agency indicated it plans to take a "hands-off approach" under the current administration, a posture that leaves the field open for states to set their own rules.
That vacuum is already being filled. Texas and Wyoming have also moved to waive rules for AI in healthcare. Lawmakers in Iowa, Idaho, and other states have introduced legislation to formally license AI medical services, many using a legislative template drafted by the Cicero Institute, a pro-AI nonprofit founded by Palantir co-founder Joe Lonsdale.
Adam Meier, the Cicero Institute's director for health policy, attributed much of the pushback against AI in medicine to economic anxiety rather than patient safety:
"Whoever goes first is going to take the slings and arrows because there's economic interests, concerns about the workforce and what that's going to mean for jobs."
That framing conveniently dismisses the substance of what the doctors are saying. The medical licensing board's March letter didn't mention jobs. It cited drug interactions and side effects, the kind of clinical concerns that exist whether a doctor's paycheck is at stake or not.
The American Medical Association has warned that "prescription renewals aren't routine checkboxes." Dr. John Whyte, described by the Washington Examiner as the AMA's CEO, put it this way: "While AI has limitless opportunity to transform medicine for the better, without physician input it also poses serious risks to patients and physicians alike."
Doctronic co-CEO Matt Pavelle struck a more optimistic note, saying the Utah partnership "enables patients, pharmacists, and physicians to work together more efficiently with measurable results that benefit the entire healthcare system." The company, founded in 2023, describes itself as HIPAA-compliant and trained on peer-reviewed medical sources.
But the measurable results so far are thin. The only performance study is the company's own unreviewed paper. Utah has released some initial data on the program, but the specifics remain sparse. And the five-member oversight board, again, with no physicians, has not publicly detailed the "numerous safeguards" it says it has implemented.
Daniel Aaron, a professor at the University of Utah law school, warned that the rush to deploy could backfire on the companies themselves:
"Companies may benefit in the short term by expanding their business models and kind of having the technology go beyond the evidence. But in the long-term, I think they risk compromising public trust and fueling backlash."
That warning applies beyond Utah. As more states explore AI in healthcare, the question of who verifies these systems, and who is accountable when something goes wrong, becomes urgent. The accountability gaps in healthcare systems are already well documented, from massive patient data breaches to federal funding disputes over Medicaid fraud enforcement.
Conservatives should welcome innovation in healthcare. The system is bloated, expensive, and often hostile to patients. If AI can genuinely reduce costs and improve access for people managing chronic conditions, that's worth exploring.
But exploring is not the same as deploying, and deploying is not the same as deploying without the input of the state's own medical licensing board. Utah didn't just move fast. It moved fast and shut out the one body specifically charged with protecting patients from unsafe prescribing.
Bressman compared the current moment to the haphazard medical standards of the early twentieth century, before medical schools, licensing boards, and national benchmarks brought order to a chaotic profession. A century of hard-won patient protections shouldn't be waived by a five-member AI panel with no medical credentials and a startup whose only performance data is self-reported.
The right answer isn't to ban AI from medicine. It's to demand the same rigor from a chatbot that we demand from a doctor, peer-reviewed evidence, independent oversight, and accountability when things go wrong.
If a human physician tried to prescribe 190 medications without a license, without board approval, and with an 80 percent accuracy rate, the state would shut the practice down. The standard shouldn't drop just because the prescriber runs on software.
