Democrats, providers and other stakeholders are also concerned about a potential lack of transparency and oversight. In their July letter, Democratic members of Congress said a lack of technical transparency about how AI-powered prior authorization solutions make decisions would make it difficult for the CMS to ensure adequate program and model oversight.

The CMS must ensure that participants, such as MA plans, are following traditional Medicare’s national and local coverage determinations, Bassano said.

“CMS should know what it is buying, or contracting for, from these model participants,” Bassano said.

Stakeholders are also concerned about financial incentives that could encourage model participants, such as commercial insurers or technology vendors, to deny necessary medical care to boost profits. Under the WISeR model, contracted entities will earn shared savings by restricting services.

The WISeR model “will reward AI bounty hunters,” Gilberg said.

“While CMMI claims the percentage will be adjusted based on certain performance measures, including provider experience, fundamentally, AI vendors will still need to deny requests in order to profit in the program,” he said.

There are also many unanswered questions about how the model will function in practice and in conjunction with other CMS policies, such as its proposed elimination of the inpatient-only list. But CMS will likely provide the “additional information in good time,” as it usually does, Bassano said.

The agency will also evaluate what’s working and what isn’t, and make changes if the WISeR model isn’t working as intended, which is “the nature of a CMS model,” Bassano said.

“There’s a learning curve on both sides whenever new models are launched, especially something as novel as this,” Bassano said. “It’s always just a matter of keeping communication open, and CMS generally wants to hear those things.”