US Medicare AI Pilot Denies 53% of Senior Care: Private AI Vendors Get 25% Cut for Every Rejected Treatment
Doctors watched seniors with spinal fractures cry in pain as private contractors cashed in on a 25% bounty for every denied procedure.
In September 2026, federal documents released through litigation by the Electronic Frontier Foundation (EFF) exposed the alarming reality behind Medicare's experimental AI prior authorization program, known as WISeR (Wasteful and Inappropriate Service Reduction). Rolled out across six states—Washington, Ohio, New Jersey, Texas, Arizona, and Oklahoma—the pilot requires doctors to secure algorithmic pre-approval from private AI vendors before performing procedures such as spinal fusions, epidural pain injections, and skin grafts on seniors.
According to the records, private contractor Virtix reviewed 6,096 prior authorization requests and denied 3,233 of them—a staggering 53 percent denial rate. Even more alarming is the program's underlying financial structure: Medicare contracts stipulate that AI vendors are paid a 25 percent bounty on the benchmark cost of every medical procedure they deny. An internal memo from the CMS Office of the Actuary explicitly acknowledged that vendors "will have an incentive to deny as many claims as possible."
While decisions were supposed to take 72 hours, patients faced delays lasting weeks, with at least one request pending for 83 days. Implementation was plagued by technical blunders—vendor Zyter failed for months to distinguish between inpatient and outpatient Medicare rules, while Innovaccer's software was so unprepared it had to temporarily auto-approve all claims to prevent backlogs. In Ohio, a healthcare provider described watching three elderly patients with spinal fractures cry at their bedside in agonizing pain waiting for AI approval, calling the system "a disgrace to the human race." Despite widespread physician outcry, federal officials still plan to expand the AI model to oncology treatments, pacemakers, and air ambulance transport.
Sources include Ars Technica, documents obtained by the Electronic Frontier Foundation (EFF), US Senate Health Committee hearing transcripts, and Stat News.
BenPig thinks that being sick is painful enough without an algorithm deciding who gets help.