Medicare pilots AI review of claims, which may carry risks

Medicare began a six-year pilot in six states that uses artificial intelligence to review prior authorization requests for 14 services; researchers say it could reduce spending but also risks delaying or denying necessary care.

Medicare has launched a six-year pilot that uses artificial intelligence to review prior authorization requests for certain services in traditional Medicare. The program, called the Wasteful and Inappropriate Service Reduction Model, began affecting beneficiaries in January 2026 in six states. It requires medical providers to request permission before performing 14 types of procedures or devices. The pilot is intended to test whether AI-supported prior authorization can reduce unnecessary care and lower spending while being monitored for inappropriate denials. Key facts: - The pilot started in January 2026 in Arizona, New Jersey, Ohio, Oklahoma, Texas and Washington and runs through December 2031. - The model requires prior authorization for 14 services and uses private technology firms that apply AI to approve or deny requests; those firms are paid a share tied to savings from denied claims. - The Centers for Medicare & Medicaid Services will monitor the pilot for inappropriate denials and evaluate results; expansion to additional services or states is possible if the evaluation finds favorable outcomes. Summary: The pilot changes how traditional Medicare handles authorization by introducing AI-supported reviews and is likely to reduce use of the targeted services, with effects on spending, provider paperwork and private contractors' roles. The health impact depends on how accurately the AI and reviewers distinguish unnecessary from necessary care. CMS will evaluate the pilot and may recommend broader implementation based on those findings.