Trump's AI Medicare Experiment Is Denying Senior Care

A federal experiment launched in January 2026 has placed an artificial intelligence system between elderly patients and the medical care their doctors say they need. The program, called WISeR, represents the first time Medicare has required prior authorization for certain categories of treatment — a significant departure from the program's six-decade history of letting physicians make clinical calls without seeking algorithmic approval. The results, documented in federal records and confirmed by congressional investigators, have been damaging.

What Is the WISeR Program and How Did It Start

Medicare has long operated without the prior authorization requirements that plague private insurance markets. Doctors treating Medicare patients have traditionally been able to order procedures, therapies, and post-acute care without submitting requests to a gatekeeper for approval. That changed when the Trump administration quietly launched WISeR in January 2026 — a pilot program that deploys AI Medicare prior authorization to evaluate certain care requests before they are approved or denied.

The administration framed the program as a tool for reducing waste and ensuring appropriate utilization of Medicare resources. But Medicare's traditional structure never required this layer of scrutiny, which means the program represents a policy reversal with real consequences for millions of seniors. Prior authorization systems in private insurance have long been criticized by physicians and patient advocates for delaying necessary care and increasing administrative burdens. Applying that same framework to Medicare — using an AI system rather than human reviewers — raised alarms almost immediately.

Early Reports of Failure: Delays, Denials, and Patient Suffering

Within weeks of WISeR's rollout, media outlets began documenting the fallout. Physicians reported technical difficulties accessing the system. Decisions that should have arrived in hours stretched into days. Patients who needed rehabilitation, home health services, or other post-acute care were left waiting while the AI processed their cases. Some denials arrived without coherent explanation, leaving doctors with no clear path to appeal or correction.

The pattern emerging from those early reports was consistent: the AI Medicare prior authorization system was not functioning as a neutral administrative tool. It was generating outcomes that neither clinicians nor patients could predict or trust. Doctors described spending hours on administrative work that previously did not exist. Patients — many of them elderly, managing multiple chronic conditions — were left in limbo. Some were in pain. Some could not access care they had already been prescribed.

These reports were not anecdotal outliers. They described a systemic breakdown across multiple care settings, from skilled nursing facilities to outpatient rehabilitation centers. The scope of the disruption suggested the program had been deployed without sufficient real-world testing or operational safeguards.

Federal Documents Reveal the Scale of the Problem

The clearest picture of WISeR's failures came not from media accounts but from the federal government's own records. The Electronic Frontier Foundation, a digital rights organization, obtained a tranche of internal documents about the program through litigation. Released earlier in September 2026, those records contained direct feedback from healthcare providers who had firsthand experience with the AI Medicare prior authorization system.

The provider testimony was stark. One healthcare worker described the program as "a disgrace to the human race." The same provider reported watching patients cry in pain while waiting for the system to issue decisions on their care. These were not abstract policy complaints — they were accounts of real suffering tied directly to algorithmic delays.

The EFF documents confirmed what physicians had been warning since January: the system was not equipped to handle the complexity of real patient cases. AI models trained on claims data and clinical guidelines cannot replicate the contextual judgment a treating physician applies when evaluating an individual patient's condition, history, and needs. When an AI Medicare prior authorization tool makes errors, those errors don't stay on a spreadsheet. They result in delayed rehabilitation, missed therapies, and patients sitting in pain waiting for a machine to reconsider.

Medical associations, including the American Medical Association, have published extensive concerns about algorithmic decision-making in insurance approvals. Their core argument is that prior authorization systems — even human-run ones — frequently delay necessary care and produce clinically inappropriate denials. Adding an AI layer to that process, without adequate oversight or appeals mechanisms, compounds those risks.

Legal and Procedural Concerns Raised by GAO

The Government Accountability Office, Congress's independent watchdog, examined how WISeR was established and released its findings in May 2026. The conclusion was direct: Trump administration officials did not follow proper procedure in setting up the program. That finding raised serious questions about the program's legality.

Federal agencies operating in the Medicare space are subject to notice-and-comment rulemaking requirements under the Administrative Procedure Act. Programs that alter coverage conditions or introduce new requirements for beneficiaries and providers are generally expected to go through a transparent regulatory process that allows public input. The GAO's determination that proper procedure was not followed suggests WISeR may have bypassed those protections entirely.

The procedural violation matters beyond bureaucratic formality. The notice-and-comment process exists to surface problems before a program launches — problems exactly like the ones WISeR created. Healthcare providers, patient advocates, and policy experts would have had the opportunity to flag the risks of deploying AI Medicare prior authorization in a population as vulnerable as Medicare beneficiaries. That input was never solicited.

Congressional Response and Calls to Shut Down the Program

Lawmakers on both sides of the aisle have demanded answers since reports of WISeR's failures began circulating. Congressional oversight committees have sought documentation from the Centers for Medicare and Medicaid Services about the program's design, its decision-making criteria, and the data behind its denials. Members of Congress have also called for the program to be suspended pending further review.

The legislative response reflects genuine alarm about what the documents revealed. When federal records show healthcare providers describing an AI program as a "disgrace" while patients cry in pain awaiting decisions, the political pressure to act becomes difficult to ignore. The combination of documented patient harm, confirmed procedural violations, and absent transparency has given critics a clear basis for demanding the program's shutdown.

Whether Congress can force that outcome depends on the administration's willingness to respond. As of late September 2026, WISeR remained operational despite the mounting evidence of its failures and the unresolved legal questions surrounding its establishment.

Broader Implications: AI in Healthcare Decision-Making

WISeR is not the only experiment in AI Medicare prior authorization, but it is the most prominent example of what happens when such systems are deployed at scale without adequate safeguards. The program's failure carries lessons that extend beyond Medicare.

AI systems applied to healthcare authorization are only as good as their training data, their design specifications, and the oversight structures built around them. When they fail, the consequences are not software errors — they are clinical harms. Seniors denied care, physicians unable to treat their patients, and a bureaucratic process that is slower and more opaque than the one it replaced.

Health policy researchers have long warned that algorithmic prior authorization risks encoding existing biases, misapplying population-level statistics to individual patients, and creating denial patterns that are difficult for anyone — provider or patient — to challenge. WISeR appears to have validated those concerns in real time.

The GAO finding that the program was set up improperly is not a procedural footnote. It is evidence that a consequential policy change affecting millions of Medicare beneficiaries was implemented without the oversight mechanisms designed to prevent exactly this kind of outcome. Until those mechanisms are restored — and until any AI Medicare prior authorization system can demonstrate it does not harm the patients it is supposed to serve — the experiment should stop.

Source: Ars Technica - All content