The battle over prior authorization is moving from fax machines to algorithms, and a lot of physicians are not happy about it. That tedious process where your insurer must sign off on a treatment before covering it was already a source of mutual frustration for patients and doctors. The idea was always to cut down on wasteful spending by steering people toward cheaper, equally effective options. In reality, it often just meant delays long enough that people gave up on care altogether, or a mountain of paperwork to appeal a denial.
Now health insurers are eyeing AI as the fix. The sales pitch is straightforward: let a model instantly approve the obvious, unambiguous claims so nobody waits two weeks for a routine green light. That part sounds great. The resistance, captured in a 2025 survey from the American Medical Association, centers on the flip side of that efficiency. A striking 61 percent of physicians are worried that AI won’t smooth out the process but will instead supercharge wrongful denials of medically necessary care, automating a ‘no’ faster than a human reviewer ever could.
The AMA isn’t calling for an outright ban. They’re pushing for guardrails—specifically demanding that insurers provide actual clinical reasoning behind a denial, not just a rubber-stamped rejection from a black box. Transparency around the algorithms themselves is another sticking point. If a model is denying your MRI based on a pattern it found in claims data, your doctor deserves to know the logic so they can fight it.
Health policy analyst Camm Epstein cut to the core of the anxiety in an email to Undark, framing the issue not as a tech problem but a moral design choice. “AI should be used to make appropriate care easier to approve,” Epstein wrote, “not necessary care easier to deny.” Meanwhile, the Trump administration is currently piloting its own AI-driven program across six states, aiming to trim what it considers unnecessary spending. Whether that experiment proves to be a scalpel that removes bureaucratic fat or just a faster blade for cutting off access to care is the multi-billion-dollar question nobody can answer yet.
