If you’re like me, you or a loved one has struggled through the process of gaining pre-approval for the medical care that your physician has recommended. Personal stories abound regarding the tribulations of patients as they go through hoops to get their health insurer to pay for certain prescription medications, medical procedures, and more.
When used judiciously, this process — known as prior authorization — serves as a check on overuse and spending on services or technologies for which there are less costly alternatives. But a large majority of physicians voice concerns about care delays, which can cause patients to abandon recommended treatments while waiting for the insurance company to verify their eligibility and confirm that the treatment is, indeed, medically necessary. Patients who are denied care may submit an appeal, but that requires more time.
AI might be able to help. With its ability to efficiently sort through vast reams of information, artificial intelligence could theoretically expedite approval of unambiguously allowable claims, thereby reducing care delays. However, AI-driven prior authorization is facing resistance, as it may increase wrongful denials of health insurance coverage. A 2025 American Medical Association survey of physicians revealed significant concern about application of AI tools, with 61 percent of doctors worrying that AI will exacerbate denials of what they deem are necessary treatments.
The AMA advocates requiring insurers to provide detailed clinical reasoning to justify denials of coverage, in addition to more transparency regarding AI algorithms.
In an email to Undark, health policy analyst Camm Epstein wrote that “AI should be used to make appropriate care easier to approve, not necessary care easier to deny.”
President Donald Trump’s administration is currently piloting a program in six states, using AI to reduce unnecessary medical spending. But it remains to be seen whether this new approach will help fix a tortuous system.
Regardless of the degree to which AI is involved, the public views prior authorization as a major burden. In Medicare Advantage — the privately run alternative to original Medicare that now enrolls roughly 55 percent of Medicare-eligible seniors and disabled people — insurers issue millions of full or partial claim denials annually based on prior authorization. Federal government reports issued in June showed that plans sometimes even reject requests for skilled nursing and rehabilitation admissions. Erecting obstacles to medically appropriate care is viewed as a particular area of concern.
Patients can request medical exemptions or appeal plan decisions, but the process is often complicated and cumbersome. NBC News reported that some patients are “stuck in prior authorization” purgatory as they “run out of time or treatment options.”
A newly released Commonwealth Fund survey finds that roughly one in five American working-age adults with private insurance reported that either they or a family member were denied insurance coverage for physician-recommended medical care in 2025. Forty-one percent of people who experienced a prior authorization denial said it delayed their care, and more than a quarter reported that their health problem worsened as a result.
“AI should be used to make appropriate care easier to approve, not necessary care easier to deny.”
The government and private insurers have tried to make improvements.
A rule issued by former President Joe Biden’s administration in 2024, for example, included reforms designed to reduce delays for patients with government-run plans while streamlining the prior authorization process for physicians. It required insurers to make certain prior authorization decisions within 72 hours for urgent requests, and seven calendar days for non-urgent requests. Per Jan. 1 of this year, these timeline requirements went into effect for most health plans in the public sector. Last year, together with insurers, the Trump administration pledged to further streamline and accelerate prior authorization processes. Private insurance companies vowed to standardize electronic requests by 2027 and to “reduce the volume of medical services subject to prior authorization” by 2026, including for common procedures like colonoscopies and cataract surgeries.
Now, the Trump administration wants to further ameliorate prior authorization protocols by expanding the use of AI.
This year, the Centers for Medicare and Medicaid Services began a demonstration project called WISeR, or Wasteful and Inappropriate Service Reduction Model. Using AI, WISeR is designed to reduce waste and fraud in original Medicare, aiming to decrease unnecessary procedures. The project runs through December 2031 in six states, and combines technologies such as machine learning with human clinical review to evaluate services CMS believes may be vulnerable to overuse, fraud, and abuse, including skin and tissue substitutes, electrical nerve stimulator implants, and knee arthroscopy for knee osteoarthritis.
Although prior authorization has been used extensively in Medicare Advantage, it has rarely been deployed in original Medicare. And this shift might not be good for patients.
A HHS Office of Inspector General memorandum published in 2022 pointed to more than one in 10 instances in which Medicare Advantage plans denied beneficiaries’ access to services even though they apparently met coverage rules. (Being denied access doesn’t mean patients can’t ever gain access. In 2024, for example, Medicare Advantage plans overturned 81 percent of denials upon appeal.)
Support Undark Magazine
Undark is a non-profit, editorially independent magazine covering the complicated and often fractious intersection of science and society. If you would like to help support our journalism, please consider making a donation. All proceeds go directly to Undark’s editorial fund.
By integrating AI into the prior authorization process, CMS says the WISeR model will “ensure timely and appropriate Medicare payment for select items and services.” But this is not how critics see it. Before WISeR was implemented, Wendell Potter, an advocate for health insurance reform and former executive at health insurer Cigna, covered the political pushback against the model on the Substack publication “HEALTH CARE un-covered.” In the same publication, Zena Wolf, a researcher with the Center for Health & Democracy, cited investigations by the Washington Post, KFF Health News, and the Seattle Times that suggest in the first few months of the year, the model has caused delays in care and denials in some instances in each of the six states where it is being piloted. And despite automated processes, there can be a high administrative burden for health care providers, which includes additional work dealing with denials.
Additionally, vendors participating in the WISeR model, who were hired to carry out AI-driven prior authorization, earn a share of what CMS calls “averted expenditures.” This could entail revenues for rejecting care requests. In turn this points to a broader discussion about longstanding concerns regarding profit-making on the basis of discouraging patients from getting medically necessary care. Several lawmakers have introduced resolutions and amendments to block funding for the WISeR model, citing threats to patient access.
Yet the Trump administration seems to be of two minds when it comes to prior authorization. As CMS expands its use in original Medicare using AI, the agency wants to lessen and streamline its use by private insurers, including Medicare Advantage plans. CMS Administrator Mehmet Oz has warned insurance company executives that they must ease the burden of prior authorization, or the federal government will impose regulation: “If you don’t do it yourselves, then we’re going to do it for you,” he told the National News Desk, a TV news program.
Regardless of the degree to which AI is involved, the public views prior authorization as a major burden.
Possibly to preempt further executive branch action or passage of laws by legislators, health plans released data recently that suggest they’re complying with administration demands. The industry-based survey reveals that between June 2025 and April 2026, requests for prior authorization declined by 11 percent. It’s unknown, however, whether the denial rate has decreased.
Responding to an industry group survey conducted last year, all responding health plans agreed with the statement, “AI or algorithms without clinician or practitioner review are not used to deny prior authorization requests that involve medical necessity or clinical considerations.” Moreover, insurers promised more transparency around clinical reasoning underlying prior authorization.
This may alleviate some of the worry about a lack of human review of decisions made by AI. But placating detractors won’t be easy.
Jared Dashevsky, a physician and founder of a media and educational platform called Healthcare Huddle, wrote that AI could “eliminate barriers, reduce administrative waste, give us more time with patients. But that’s not what’s being built.” Instead, he says, there’s an “arms race to deny faster and appeal faster. More automation of a broken system that shouldn’t exist in its current form.”