How AI May Soon Prescribe Mental Health Medication in the U.S.

Artificial intelligence is steadily moving into areas once handled only by physicians. One of the latest developments involves mental health treatment, where AI systems may soon assist with prescription renewals. A new program aims to reduce long wait times, high appointment costs, and limited access to psychiatric care.

The idea is straightforward: technology handles routine prescription renewals while doctors step in when concerns arise. A startup called Legion Health is preparing to launch what it describes as the first mental health platform authorized to allow artificial intelligence to prescribe certain psychiatric medications.

The initiative begins in Utah and could expand nationwide if the early rollout proves safe and effective.

A New Approach to Mental Health Prescriptions

Linkedin | Yash M. Patel | Arthur MacWaters, Yash Patel, and Daniel Wilson, founded Legion Health to bridge medical access gaps using AI.

For many patients, renewing mental health medication can be frustrating. Scheduling an appointment may take weeks, and copay costs can quickly add up. Even individuals with insurance often pay hundreds of dollars simply to receive a routine refill.

Starting next month, the system created by Legion Health aims to change that process. Patients will be able to request prescription renewals through an AI-powered platform for a monthly subscription fee of about $20. The company, backed by Y Combinator and funded with roughly $7 million since its launch in 2021, plans to make routine medication management faster and more accessible.

The rollout will begin only in Utah, where healthcare shortages have created long delays for many patients seeking psychiatric care. Expansion into additional states is expected once the pilot program gathers sufficient data.

Team Behind the AI Mental Health Platform

The company was founded by three entrepreneurs: Arthur MacWaters, Yash Patel, and Daniel Wilson. All three founders are currently 29 or 30 years old and launched the company with the belief that artificial intelligence could help address gaps in medical access.

According to cofounder Arthur MacWaters, the long-term goal is not to replace physicians but to combine technology with clinical oversight. He explained that the vision involves “AI + doctors + clinic in the loop,” allowing artificial intelligence to safely handle specific medical tasks while doctors remain involved when needed.

MacWaters also described the broader opportunity for AI in medicine, noting that the concept of an “AI doctor” could become one of the most valuable sectors worldwide.

How the AI Prescription System Works

The program is intentionally limited in its early phase. The AI will not diagnose patients or prescribe new medications. Instead, it focuses strictly on renewing medications that were previously prescribed by a licensed doctor.

The system currently handles lower-risk psychiatric maintenance medications, including commonly used antidepressants and sleep aids such as SSRIs, Wellbutrin, Trazodone, and Mirtazapine.

Patients must actively choose to use the AI system. Once enrolled, they can access the platform through a mobile app or a web browser. The interface clearly informs users that they are interacting with an AI agent rather than a human clinician.

After verifying identity and receiving patient consent, the AI performs a short safety evaluation that lasts roughly two minutes. During this review, the system checks several important factors, including medication interactions, potential side effects, and psychiatric warning signs.

If the AI detects any potential risk, a human doctor immediately takes over the case. Patients can also request a human review at any time.

Why Access to Care Is Limited in Utah

Freepik | Utah’s pervasive healthcare shortage leaves patients facing extreme travel, long waits, and high costs for routine care.

The program’s first launch location reflects a larger issue in American healthcare. Every one of the 29 counties in Utah has been classified as a health professional shortage area. This designation indicates a severe lack of medical providers across the state.

As a result, many residents travel long distances to meet with specialists or wait weeks just to schedule a simple prescription renewal. In some cases, patients must drive hours and still face appointment delays of two weeks to two months.

Cofounder Daniel Wilson pointed to this challenge while discussing the project. He explained that patients often pay around $300 out of pocket for visits that only result in a routine medication refill.

Developers designed the program with several safety checkpoints before allowing the AI to operate independently. The rollout follows a phased approach that includes human oversight throughout the early stages.

The first 250 prescriptions processed by the system will require direct supervision from doctors. After that phase, the next 1,000 prescriptions will undergo post-evaluation reviews by medical professionals. Only after those stages will the AI begin operating autonomously.

This gradual testing strategy aims to build trust while identifying potential issues before expanding the technology more widely.

The Role of AI in Medical Decision-Making

Advocates of AI in healthcare argue that technology can reduce common human limitations. Doctors often see dozens of patients per day while working with outdated software systems. According to Arthur MacWaters, some medical record platforms feel like “this Windows 1994 stuff,” highlighting the aging infrastructure used in many clinics.

Artificial intelligence systems offer several advantages in comparison. They do not become fatigued, and they can review thousands of pages of medical records in seconds. This ability allows AI tools to quickly identify potential drug interactions that busy physicians might overlook.

MacWaters emphasized that the larger issue is capacity. In his view, there simply are not enough human doctors worldwide to meet the growing demand for healthcare. Artificial intelligence may help close that gap.

The Mixed History of Tech in Healthcare

Technology companies have attempted to reshape healthcare before, often with mixed outcomes. One of the most well-known failures involved Theranos, a startup that collapsed after fraudulent claims about its blood-testing technology.

Other large initiatives have struggled as well. For example, Amazon previously launched a healthcare venture called Haven that ultimately shut down.

Still, some digital health platforms have found lasting success. Telemedicine provider Teladoc Health has built a large remote-care network that connects patients with doctors through virtual visits.

These contrasting outcomes highlight how difficult healthcare innovation can be, especially when new technologies attempt to change long-standing medical systems.

Utah’s Unique Approach to AI Regulation

Part of the reason the pilot program exists in Utah involves the state’s regulatory approach to artificial intelligence. Officials have created regulatory “sandboxes,” which allow companies to test new technologies under temporary exemptions from certain rules.

This model encourages experimentation while still maintaining government oversight. According to Margaret Woolley Bussee, Executive Director of the Utah Department of Commerce, the state wants to avoid extreme positions on AI.

Bussee described the strategy as a middle path between unrestricted adoption and complete prohibition. Policymakers aim to study how the technology performs in real-world scenarios before implementing broader regulations.

governor.utah.gov | Executive Director Bussee defines Utah’s AI strategy as a cautious blend of open experimentation and active state supervision.

Not every state shares the same outlook on artificial intelligence in medicine. Some regions have begun considering strict restrictions on AI-generated health advice.

For example, lawmakers in New York have proposed legislation that could block AI systems, including ChatGPT, from responding to health-related questions. The proposal would limit even basic discussions about drug interactions or symptoms.

These contrasting policies may create very different healthcare environments across the country. In the near future, traveling across state lines could mean entering an entirely different regulatory system—one where AI acts as a healthcare assistant or where it is not allowed to participate in medical conversations at all.

The Future of AI in Healthcare

Supporters of artificial intelligence in medicine expect the technology to become a regular part of patient care. Developers believe AI systems could assist with tasks such as reviewing medical records, monitoring patient data, and supporting treatment decisions.

Arthur MacWaters has expressed confidence in this shift, suggesting that within the next five years, most patients may have AI tools working in the background to support their healthcare needs.

Early initiatives already show how this model might function. A pilot program from Legion Health in Utah is testing a system where artificial intelligence handles routine prescription renewals for certain psychiatric medications while doctors remain responsible for complex or high-risk cases.

At the same time, government policies will heavily influence how quickly these technologies spread. Some states are experimenting with controlled AI testing environments, while others are considering tighter restrictions on how AI can be used in medical settings.

As these policy approaches develop, healthcare systems across the United States may begin to look very different depending on local regulations and the results of early pilot programs.