This bill mandates that Medicare Advantage plans ensure all AI-driven prior authorization denials are reviewed and approved by qualified physicians to protect patients from automated decision-making.
Herbert Conaway
Representative
NJ-3
The Protecting Patients from Automated Denials Act mandates that Medicare Advantage plans cannot deny prior authorization requests based solely on artificial intelligence. The bill requires that all such denials be reviewed and signed by a qualified physician who exercises independent medical judgment. Additionally, it establishes strict transparency, record-keeping, and reporting requirements to ensure AI is used only for administrative purposes and remains subject to federal oversight.
Medicare Advantage Plans Face New Human-in-the-Loop Rules: AI Cannot Deny Your Care Without a Doctor’s Signature Starting in 2027
If you or your parents are on a Medicare Advantage (MA) plan, you know the 'prior authorization' dance: your doctor says you need a scan or a specialist, but the insurance company has to give the green light first. Increasingly, those green lights—and the 'no's'—are being handled by algorithms. This bill, the Protecting Patients from Automated Denials Act, steps in to ensure a computer doesn't have the final word on your health. Starting January 1, 2027, MA plans are prohibited from denying a medical request based solely on AI output. Every single denial must be reviewed and signed by a human physician who holds an active license and specializes in the area of medicine they are reviewing.
Putting the 'Human' Back in Healthcare
The bill creates a paper trail for every 'no.' Before an insurance company can reject a claim using AI, a qualified physician reviewer must sign an attestation—basically a formal promise—stating that they used their own independent medical judgment and that the software was only used for administrative help. For example, if a software developer on an MA plan needs a specific knee surgery, a computer algorithm can't just flag the request as 'not standard' and trigger an automatic denial. A board-certified orthopedic specialist would have to look at the file, agree with the denial, and put their National Provider Identifier (NPI) number on the line. You and your doctor will also get a copy of this signed statement and a direct line to talk to that reviewer about why they made that call.
Receipts and Accountability
To make sure insurance companies aren't just rubber-stamping AI decisions, the bill mandates serious transparency. Every 90 days, starting in early 2027, plans have to send a report to the Department of Health and Human Services detailing every AI-related denial and the algorithms they used. They also have to keep these records for 10 years. This gives the government the 'receipts' needed to perform audits. If a plan is found to be using a 'black box' algorithm that consistently denies valid claims, the Secretary of Health and Human Services has the authority to dig into their internal policies, interview employees, and check if those 'independent' doctors are actually doing their jobs or just clicking 'agree' on every AI prompt.
The Cost of Compliance
While this is a win for patient rights, it adds a significant layer of bureaucracy for the insurance companies managing Medicare Advantage. Hiring board-certified specialists to manually review every AI-flagged denial is more expensive than letting a program run the numbers. For the busy professional or the trade worker, this might mean that while your care is more protected, the administrative costs for these plans could climb. However, the bill is betting that the price of human oversight is worth it to prevent patients from being trapped in an automated loop where a computer error stands between them and their recovery.