An artificial intelligence system is now making treatment decisions for Medicare patients in six U.S. states, and the results are already harming people. The program, called the Wasteful and Inappropriate Service Reduction (WISeR) Model, requires doctors to submit their clinical reasoning to an AI system before patients can receive certain procedures. What was promised as a 72-hour approval process has instead created delays lasting weeks, leaving elderly patients suffering while waiting for decisions made by algorithms they cannot see or challenge.

How Is AI Currently Making Healthcare Decisions for Medicare Patients?

The WISeR pilot program operates in Washington, Arizona, New Jersey, Ohio, Oklahoma, and Texas. Under this system, doctors must upload their reasoning for 15 specific treatments to an online portal, where AI provided by third-party tech companies assesses whether the procedure should be covered by Medicare. If the AI denies a treatment, a human clinician at the tech company is supposed to manually review that decision.

The treatments covered by the program range from epidural steroid injections for pain management to skin substitutes and knee arthroscopy. These procedures were selected because they are either often overused or have historically had a greater risk of fraud, waste, or abuse, according to the Centers for Medicare and Medicaid Services (CMS), which oversees the scheme.

Keith Magnuson, an 83-year-old from Seattle, experienced the system firsthand. His doctor recommended a minimally invasive lumbar decompression procedure to address his debilitating back pain from lumbar spinal stenosis. The procedure itself would have been covered by Medicare, but the epidural steroid injection component was denied by the AI system. After his doctor requested approval twice more with additional documentation, Magnuson discovered that no human had made the decision at all. "I was outraged," Magnuson said. "I was like, wait a minute, it's not even another person at the other end? It's AI? It's a bot?".

What Problems Are Doctors and Patients Experiencing With the AI System?

The WISeR program has encountered significant operational challenges since its rollout. Treatments that once took a day to authorize are now taking weeks or longer. Jeb Shepard, director of policy at the Washington State Medical Association, explained the disconnect between promises and reality: "This was sold to the physician community as: 'We're going to turn this around in 72 hours, it's super fast,' and instead we've had people waiting four weeks or beyond. And this is for a patient population that's elderly; if they don't get timely care, their conditions deteriorate pretty rapidly compared to someone who is middle-aged or younger".

Beyond delays, doctors report additional problems with the system:

• Technical Glitches: Hospitals and medical practices face unexpected denials of care and technical malfunctions that require extra paperwork and often necessitate calling patients back in for additional appointments.
• Lack of Transparency: While some tech companies are responsive in explaining why care has been denied, others have been difficult to reach by physicians or have struggled to provide clear explanations for treatment rejections.
• Black Box Decision-Making: Dr. Jeff Marr, a health economist and assistant professor at Brown University's School of Public Health, described the system as "a complete black box." He noted that "how these models work, what information they're looking at, what data they're trained on" remains largely opaque to researchers and the public.

Dr. Steve Aydin, a pain doctor at Kayal Orthopaedic Center in New Jersey, questioned the fundamental premise of using AI for treatment approvals. "I don't think AI has a place in making approvals on whether a treatment is appropriate or not," Aydin stated. "A clinician is making a decision based on information that's synthesized from an evaluation, a conversation, a history and a physical exam; the AI doesn't know what the patient is feeling or going through".

Why Was This AI System Implemented Without Broader Consultation?

The Trump administration implemented the WISeR program with minimal input from the medical community. Michelle Mello, a professor of health policy and law at Stanford University, explained that the administration was able to bypass the typical regulatory process because the scheme has been classified as voluntary. This means it does not have to go through formal rulemaking that would normally require Congressional approval. However, the voluntary designation applies only to the tech companies providing the AI tools; doctors who do not participate do not receive Medicare funding, making non-participation effectively impossible for most practices.

The program was rolled out as part of the Trump administration's broader push to embrace AI as part of efforts to achieve "global dominance" in the sector, which includes reducing barriers to AI use within healthcare. The stated goal is to protect U.S. taxpayers from "wasteful" spending within Medicare and prevent patients from receiving "unnecessary" procedures. The CMS cited a report from the Medicare Payment Advisory Commission estimating that up to $5.8 billion spent on Medicare in 2022 was on "services with minimal benefit".

What Are the Regulatory Gaps That Allowed This to Happen?

The WISeR program operates in a regulatory gray zone. While parts of AI are regulated under the U.S. Food and Drug Administration (FDA), the relevant statute dates from 1976, well before the advent of advanced computing or artificial intelligence. Mello noted that "the FDA has, over the years, issued opinions stating that it thinks certain kinds of software can be considered a medical device, and have to go through FDA review." However, in practice, around 9 percent of AI-related software undergoes FDA review, leaving most AI healthcare tools largely unregulated.

Mello

There are very few specific laws around the use of AI in U.S. healthcare decision-making. This regulatory vacuum allowed the administration to implement a system that directly affects millions of Medicare beneficiaries without the typical safeguards or public input that would accompany a formal regulatory change.

How Are Medical Schools Preparing Future Doctors for an AI-Driven Healthcare System?

While the WISeR program reveals the dangers of poorly implemented AI in healthcare, medical schools are taking steps to prepare the next generation of physicians to work alongside these tools more thoughtfully. Over 80 percent of physicians now report using artificial intelligence in a professional context, according to the American Medical Association (AMA) 2026 Physician Survey on Augmented Intelligence. This represents more than double the 38 percent who reported using AI in 2023.

The Association of American Medical Colleges (AAMC) has documented a significant increase in AI integration across medical school programs between 2023 and 2024, with ongoing guidance to ensure students graduate with the knowledge to engage these tools ethically and effectively. The goal is to produce physicians who understand how AI models work, where they are reliable, and how to interpret their outputs critically.

Medical schools are training students to recognize when an algorithm may be performing poorly and to understand that physician judgment carries legal and moral weight no model can absorb. AI for future doctors means treating these systems as decision support tools, not decision-makers. This includes immersive simulation tools powered by large language models that respond dynamically to a student's questions, tone, and clinical reasoning in real time, creating practice environments where learners can rehearse high-stakes scenarios without risk to actual patients.

What Role Should AI Play in Healthcare Decision-Making?

Beyond the WISeR controversy, researchers and clinicians are exploring how AI can genuinely improve patient care when implemented thoughtfully. Dr. Michael Matheny, professor of medicine and biomedical informatics at Vanderbilt University, believes that AI, used as a tool, can improve the patient-practitioner relationship and healthcare overall. His work focuses on two key areas: implanted medical product surveillance and improved patient medical records.

Matheny noted that "with increasing access to health data, and advances we and others have made in the use of AI in medical device safety, we are able to more efficiently identify medical product signals that might harm patients." The World Health Organization estimates there are about 2 million types of implantable medical devices on the market, from pacemakers to artificial knees to cochlear implants. AI can help identify safety issues across these devices more efficiently than traditional surveillance methods.

Matheny

Another area where AI shows promise is in helping clinicians manage the overwhelming volume of information in patient medical records. Matheny explained: "Human cognition is limited, and a provider cannot be expected to manage terabytes of information on every patient going back decades into their medical history." He and his colleagues are partnering with the Veterans Affairs system on a project to help inpatient care teams leverage AI in the diagnosis of severe kidney injury by summarizing years of patient clinical information and helping providers consider different causes and etiologies.

Matheny

The key difference between these applications and the WISeR system is that they are designed to support physician decision-making rather than replace it. When AI is used to help doctors synthesize complex information or identify patterns they might miss, it can enhance care. When AI is used to deny treatment without transparent reasoning or physician oversight, it undermines the trust and clinical judgment that are essential to medicine.