On July 16, the Senate voted 46 to 50 against a resolution that would have killed a new Medicare pilot, leaving it fully in place [1]. The pilot, called WISeR, does something original Medicare has almost never done: it applies prior authorization - the insurer's 'may I' step before certain care - to the traditional, government-run program, with algorithms helping decide [2].
Prior authorization is familiar in Medicare Advantage, the privately run alternative, where it has long been a source of complaints. In original Medicare - the fee-for-service program most seniors picture - it has been rare to nonexistent [2]. WISeR imports it: a six-year test, in six states (Arizona, New Jersey, Ohio, Oklahoma, Texas, and Washington), for a set of services CMS considers prone to overuse [1][2].
The role of the artificial intelligence is where careful reading matters. WISeR uses AI and other technology to flag prior-authorization requests for review; under the model's own rules, a licensed human clinician must review every denial before it is issued, and the algorithm cannot deny care on its own [3]. That distinction cuts against the sharpest headlines, which suggest a machine turning seniors away. The documented worry is subtler: the American Hospital Association warns that health plans often over-rely on the AI's recommendations, so the human review that is supposed to check the machine can become a rubber stamp [4].
The record worth completing is what the vote preserved: not a robot denying a grandmother's care, but a real and consequential expansion - prior authorization, AI-assisted, entering the one part of Medicare that had mostly been free of it, with human reviewers whose independence hospitals are already questioning [1][3][4]. The Senate had a chance to stop the experiment before it scales. It declined [1].
