Will AI Fix Prior Authorization or Make It Worse?

Prior authorization

It's a process that has long frustrated patients and physicians alike. What was designed as a simple check to prevent the overuse of medical services has turned into a constant source of friction. Patients abandon necessary care. When insurers require approval before paying for a recommended treatment or medication, the resulting delays can lead people to give up entirely, so the push to automate this system through artificial intelligence is gaining momentum as healthcare looks for ways to modernize. But will this shift actually solve the problem or just make it faster to say no?

The rise of automated denials

Advocates for change argue that artificial intelligence can process information far faster than human reviewers. It's incredibly fast. By clearing unambiguous claims in seconds, the technology could theoretically reduce the waiting period that currently leaves so many patients in limbo, but there's a clear divide between potential efficiency and the reality of implementation. So recent data suggests the push for automation is triggering alarm among those on the front lines of patient care.

Physician concerns over technology

A 2025 American Medical Association survey found that 61 percent of doctors fear these tools will cause more denials of medically necessary treatments. But the real problem is clinical nuance. When an algorithm makes the call, it's easy for the complexity of a patient's case to get lost in the code, and without transparency doctors can't understand why a specific treatment was rejected. They're left unable to effectively advocate for their patients.

• One in five working-age adults with private insurance reported a denial for recommended care in 2025.
  Market Context: According to IQVIA, 70% of commercially insured patients were initially denied coverage for at least one newly prescribed brand medicine in 2024.
• Forty-one percent of patients who faced a denial experienced a delay in receiving treatment.
• More than 25 percent of patients who were denied care saw their health problems worsen.
• Medicare Advantage plans overturned 81 percent of denials upon appeal in 2024.

A government experiment

The federal government is currently testing its own version of this technology through a project known as WISeR. It's operating in six states through December 2031. Critics argue this model introduces a dangerous incentive structure. The initiative aims to reduce wasteful spending in original Medicare by targeting specific areas for review, such as knee arthroscopy and skin substitutes, and it combines machine learning with human oversight. That's a risk.

AI should be used to make appropriate care easier to approve, not necessary care easier to deny.

Health policy analyst Camm Epstein captured that sentiment. This is the core of the project's central tension, as critics worry profit is becoming the primary driver of patient access because vendors hired to manage these denials often earn a portion of the money saved by blocking claims. But some lawmakers are now moving to block funding for the model. It's a direct challenge to the program's future, and they're citing threats to the quality of care we've all come to expect from these initiatives. They can't ignore the mounting pressure.

The push for accountability

The government expands AI use in some areas. But it's simultaneously pressuring private insurers to behave better, as CMS Administrator Mehmet Oz has taken a firm stance with industry executives and suggested that if insurers don't improve their own prior authorization processes, federal regulations will force their hand. So the industry has responded by claiming an 11 percent decline in prior authorization requests between June 2025 and April 2026. Yet the real question remains. Whether that decrease actually translates into fewer denials is still open.

The danger of an arms race

The current environment feels like a contest of speed rather than a commitment to health outcomes. But as more companies adopt automated systems, the focus shifts toward who can process the most denials in the shortest amount of time, and that's a troubling reality for many. So automation, in its current form, appears to be cementing a system that many argue shouldn't exist as it does today. It's a troubling shift.

What happens next

Patients remain caught in the middle. But while the promise of faster approval times is enticing, the history of insurance claims suggests that administrative shortcuts often come at a cost to the sick. It's a tough spot. Whether the coming years bring a more streamlined experience or a more difficult one depends on how much transparency is demanded of these new algorithms. Until the system prioritizes patient needs over the speed of claim rejections, the promise of a better process will likely remain out of reach. So don't hold your breath.