First AI drug prescriber in US sparks concerns

Experts have raised concerns over the launch of the first artificial intelligence (AI) drug prescribing service in the US, warning that, unlike doctors, AI prescribers face no licensing or competency requirements.

In January 2026, the state of Utah announced a partnership with the AI company Doctronic to prescribe medications without physician involvement – a first in the US.

The system, which was quietly rolled out as a pilot in December 2025, is designed to handle prescription renewals across nearly 200 drugs, including corticosteroids, statins, antidepressants, hormones, and anticlotting agents.

According to the Utah Office of Artificial Intelligence Policy, the AI service can process 30, 60 and 90-day renewals for applicable medications that have already been prescribed by a licensed provider. However, the service cannot issue new prescriptions or handle refills for controlled or addictive substances.

Matt Pavelle, co-chief executive officer of Doctronic, said: ‘This is a major milestone to demonstrate how AI can improve access to care and health outcomes… This partnership with Utah enables patients, pharmacists, and physicians to work together more efficiently, with measurable results that benefit the entire healthcare system.

In an opinion piece, published in JAMA, law professors Daniel G Aaron (University of Utah) and Christopher Robertson (Boston University) argued: ‘Although physicians must demonstrate competence through rigorous training and board certification, AI prescribers face no equivalent gatekeeper.

Compounding these concerns, the authors say that AI companies rarely reveal exactly how their software arrives at medical decisions. In the case of Utah, Doctronic only offered vague reassurances they said, including that its ‘trained on established medical protocols,’ and that its algorithm continues to evolve through ‘feedback loops.’

The authors also highlighted that the data used to support Doctronic’s approval in Utah was unblinded, produced by company-affiliated authors and has not been published in a peer-reviewed publication.

Advocates for AI prescribers argue that such services are essential given the growing shortage of primary care physicians. They also say AI could reduce prescription errors and free up doctors for more complex work.

The authors agree there are potential benefits as medication non-adherence alone costs the US economy over $500 billion annually. In the UK, the NHS in England spends more than £19 billion on medicines. Given that prescribed medicine non-adherence is typically 50%, some have suggested that £9.5bn worth of drugs were either not taken as recommended or not taken at all.  By lowering barriers to access, AI prescribers could improve both accessibility and adherence.

Nevertheless, the authors underscore the legal grey area surrounding the AI prescriber. US federal law requires prescriptions to come from licensed practitioners, which AI is not. While Utah bypassed this via a regulatory sandbox approach – waiving state laws for novel businesses – the authors warn it could have negative consequences in the future.

They said: ‘Without the licensed practitioner’s prescription, drugs become misbranded, making their sale a crime. It is unclear how the federal government will apply this law to entities unaffiliated with the drug manufacturer.’

Due to these legal ambiguities, the authors note that patients harmed by AI prescribing in future will face an uphill legal battle, as courts haven't established whether AI is held to the same standard of care as a doctor.

They conclude: ‘Incorporating AI into modern health care must be evidence based and responsible. Physicians and health systems should insist that AI technologies not be allowed to bypass long-standing legal guardrails governing medical products.’

In the UK, the General Pharmaceutical Council (GPhC) recently published a new position statement on the use of AI in pharmacy, alongside supporting advice on the use of AI in education and training, and in revalidation.

It states that pharmacists, pharmacy technicians and pharmacy owners must continue to meet the GPhC’s standards at all times, including when using AI.

It clarifies that pharmacy professionals remain personally accountable for their decisions and actions and must always use their professional judgement when AI tools are involved.

The statement also emphasises that pharmacists and pharmacy technicians should be transparent about their use of AI tools, protect patient confidentiality, and explain to patients how AI is being used in their care, and any associated risks or benefits.

Kathie Cashell, chief executive of the GPhC, said: ‘Artificial intelligence has the potential to support improvements in pharmacy services and patient care, but it also brings new challenges. We want to support everyone working in pharmacy to use AI effectively to improve patient care, at the same time as protecting patient safety.’

In March 2024, it was reported that a new AI tool, called DrugGPT, was being developed at Oxford University to reduce medication errors made by prescribers.