Government Pilots AI for Medicare Prior Authorization, but Doctors Fear More Denials

The Trump administration is testing an AI system to streamline prior authorization in original Medicare, a process notorious for delaying care. The pilot, called WISeR, is running in six states and uses machine learning to flag services the Centers for Medicare and Medicaid Services suspects are prone to overuse or fraud. It’s one of the biggest moves yet to inject AI into the bureaucratic bottleneck between a doctor’s recommendation and an insurer’s approval.

The stakes are high. Prior authorization is already a major headache for patients in Medicare Advantage, where federal reports show plans sometimes reject claims for skilled nursing and rehab even when patients qualify. Now CMS is bringing a version of that scrutiny to original Medicare, which has rarely used it. The agency says the WISeR model will ensure timely and appropriate payment. Critics say it’s just adding another layer of automated denial.

Doctors are skeptical. A recent American Medical Association survey found 61 percent of physicians worry AI will increase wrongful denials of treatments they consider necessary. Health policy analyst Camm Epstein put it bluntly in an email to Undark: the technology should make appropriate care easier to approve, not necessary care easier to deny. Early reports from the pilot states suggest some delays and denials are already happening.

The administration argues the system combines AI with human clinical review to target waste and abuse in areas like knee arthroscopy and nerve stimulator implants. But the fundamental tension remains: efficiency for insurers can mean obstruction for patients. With the pilot running through 2031, builders and founders in health tech should watch closely—this is a live test of whether AI can untangle healthcare bureaucracy or just make it faster at saying no.