The Centers for Medicare and Medicaid Services is testing an AI-assisted prior authorization program in original Medicare, a move that could reshape how some patients get approval for care before it is paid for. CMS says the model is meant to cut waste, fraud, and unnecessary procedures. Physicians, patient advocates, and some lawmakers see a more familiar machine: a denial system with better tooling.
The demonstration, called WISeR, short for Wasteful and Inappropriate Service Reduction Model, runs through December 2031 in six states. CMS says it combines technologies including machine learning with human clinical review to assess selected services the agency considers vulnerable to overuse, fraud, or abuse. The list includes skin and tissue substitutes, electrical nerve stimulator implants, and knee arthroscopy for knee osteoarthritis.
Prior authorization is already a major part of Medicare Advantage, the privately run alternative to original Medicare. KFF says Medicare Advantage now covers roughly 55 percent of Medicare-eligible seniors and disabled people, and insurers make millions of full or partial prior authorization denials each year. Original Medicare has used the tool far less often, which is why WISeR has drawn attention beyond the usual health-policy trench warfare.
What AI changes, and what it does not
Prior authorization asks a doctor or other provider to get an insurer or payer to approve a service before it is covered. In theory, it can block low-value care or steer patients toward cheaper alternatives. In practice, doctors say it delays treatment, adds paperwork, and can push patients to abandon care their clinicians recommended.
AI could speed approvals for requests that clearly meet coverage rules. That is the optimistic version. The darker version is that automation helps payers reject more claims, faster, while making appeals harder for patients and clinicians who have to fight the output.
The American Medical Association’s 2025 physician survey found that 61 percent of doctors were worried AI would worsen denials for treatments they consider necessary. The AMA has called for insurers to give detailed clinical reasoning when they deny coverage and to provide more transparency about AI algorithms. In an email to Undark, health policy analyst Camm Epstein wrote that AI should make appropriate care easier to approve, not necessary care easier to deny.
The concern is not theoretical. A 2022 memorandum from the Department of Health and Human Services Office of Inspector General found that Medicare Advantage plans denied access to services in more than one in 10 reviewed cases even though the requests apparently met coverage rules. KFF reported that Medicare Advantage plans overturned 81 percent of denials on appeal in 2024, which says something unflattering about the first pass.
Incentives under scrutiny
CMS says WISeR will help ensure timely and appropriate Medicare payment for selected items and services. Critics are focused on how the program pays its vendors. According to Reed Smith’s summary of the model, companies hired to operate AI-driven review can receive a share of what CMS calls “averted expenditures.” That structure has raised fears that vendors may profit when care is not approved.
Wendell Potter, a health insurance reform advocate and former Cigna executive, has written about political opposition to WISeR in his publication HEALTH CARE un-covered. In the same publication, Zena Wolf of the Center for Health & Democracy cited investigations by the Washington Post, KFF Health News, and the Seattle Times suggesting the model caused delays and denials in some cases in the six pilot states during its first months.
Several lawmakers have introduced resolutions and amendments aimed at blocking WISeR funding, citing risks to patient access. Meanwhile, CMS Administrator Mehmet Oz has pushed private insurers to reduce prior authorization burdens, warning executives in an interview with the National News Desk that if they do not act, the federal government will.
Insurers say they are changing. An industry survey reported an 11 percent drop in prior authorization requests between June 2025 and April 2026, though KFF has noted that public data do not yet show whether denial rates fell. Health plans also told an industry group survey that they do not use AI or algorithms without clinician or practitioner review to deny requests involving medical necessity or clinical considerations.
That promise leaves the central question intact. AI can process forms quickly. It cannot decide whether a health system uses that speed to clear care or to build a faster gate.
This story draws on original reporting from Ars Technica.
