Seniors just got another reason to hate AI.

The government is piloting a federal program that uses artificial intelligence to help decide what care to approve or deny. Medicare is included in the test program — dubbed the Wasteful and Inappropriate Service Reduction (WISeR).

These AI models are being used at select clinics in Texas, New Jersey, Oklahoma, Ohio, Washington and Arizona. After launching in January 2026, a lawsuit was filed against the Centers for Medicare and Medicaid Services (CMS), revealing a slew of issues within the first few months. 

AI models across six states are denying Medicare patients care Clement Coetzee/peopleimages.com – stock.adobe.com

The program’s administrator, Dr. Mehmet Oz, said WISeR would use AI to review Medicare requests for medical treatments and services. This marks a major change, as original Medicare beneficiaries rarely needed approval from their insurance to receive certain services.

“Tasking an algorithm with making determinations about treatment can create unwarranted — and even discriminatory — delays or denials of necessary medical care,” EFF’s director of AI and access-to-knowledge legal projects, Kit Walsh, said in a press release. 

“Given these serious risks, the public requires transparency that it hasn’t gotten. We’re suing to get badly needed answers about how Medicare’s AI experiment works,” Walsh added.

While the government’s goal is to reduce unnecessary medical services and any wasteful spending, healthcare experts, providers and lawmakers are worried over the program’s algorithm as there’s little information on how it’s being used — including the training data it relies on. 

Additionally, there are too many unknowns — leaving folks in the six states vulnerable and in limbo for necessary medical care. 

The EFF originally filed the lawsuit to force the government to disclose records showing how — or if — the systems were tested for accuracy, bias and AI-generated errors. Additionally, it’s unclear whether or not the government audited or monitored the performance.

Medicare patients in Texas were denied treatment by AI at a higher rate than humans G_stockerthailand – stock.adobe.com

The suit also found a couple of other important issues — a major contrast in approval numbers and a financial incentive to deny care to those people.

On page five of the 10-page lawsuit, the WISeR Model’s AI systems were “denying care at disproportionately high rates.” For example, the AI models in Texas approved care for just 62% of folks, while humans approved 84%. 

Page four of the filing revealed that contracted companies had financial incentives to deny prior approval — going against the best interests of patients. This is because vendors are paid on the volume of healthcare services that they deny and are entitled to nearly 20% of those saved costs.

“In just a few months since the launch, harms have already materialized. Physicians and other health care providers have reported unusually high denial rates, delays in care approval, technical failures, communication gaps, and administrative strain,” the lawsuit reads.

According to The Columbus Dispatch, CMS is aiming to release a six-month report that will detail WISeR’s performance. A spokesperson for CMS told the outlet that the report will include metrics and how the models are faring comparatively.

An exact date for when the report will be dropped was not given.