In 2025 alone, doctors and legislators across the country have already introduced more than 100 laws to reform advance permits. The tedious approval process is the main cause of burnout among physicians, and changing it is a top priority for the American Medical Association (AMA). According to a 2024 survey, 93% of doctors said that prior approval led to patient harm.
Leaders at Health Technology Arena believe that advance approval is ripe for artificial intelligence (AI)-driven transformation. The problem is that most doctors are not on board.
In February, the AMA reported that 61% of doctors believe that using AI with previous permits would increase the rate of rejection and delay patient care. Additionally, medical associations in at least five states have introduced legislation intended to limit the use of AI in patient decisions.
Innovators promise speed and better access, but doctors say insurance companies can easily use technology to further tax approvals and appeals. Medscape Medical News spoke to previous licensed experts, including frontline doctors. They all agree that the AI ​​may change the game, but only if the pre-authorization has obtained a transformation. If the process is not reformed to be more transparent and patient-centric, AI simply makes the flawed system work faster.
About Medscape Data
Medscape continuously explores physicians and other healthcare professionals on key practice challenges and current issues, creating highly impactful analyses. For example, Medscape Doctors and Pre-Authorized Authorization Report 2024 discovered:
Sixty-six percent of doctors said they spent a little or far more time on previous permits than before.89% of doctors say that too many drugs and services require prior approval.86% said it would interfere with patient care.19% support greater automation.
Can AI fix the system?
“The risk of stacking AI on top of a broken system offers a lot in terms of efficiency, but introducing more efficiency and speed into that system speeds up the rate at which you do bad things.
It is unclear how much AI is already being used in previous approvals. A new survey of 93 health insurance companies, announced at the National Insurance Commission (NAIC) meeting in March 2025, reported that 92% of people have been considering use, plans for use, or AI and machine learning. Preliminary results were not explicit about how health insurance companies are adopting AI.
If they use algorithms to collect relevant data from medical records more quickly, then it’s a use case where no one has any issues, Killerea said. However, algorithms that make decisions about claims and reject care are problematic. Cigna and UnitedHealth are already facing class action lawsuits after investigations were shown using AI to deny the claim without physician supervision.
According to Leslie Lenert (MD), a primary care physician studying computing at South Carolina Medical School in Charleston, companies used rule-based AI to compare claims to insurance companies’ standards. These types of if/then algorithms are fragile to the complex inferences needed to evaluate claims, Lenert said. As a result, thousands of claims were denied at a time without meaningful doctor reviews.
The Myth of Anti-Technology Doctors
Doctors have rallyed to delay the use of AI in previous permits, but contrary to popular beliefs and some headlines, it’s not because they are anti-technical.
Most doctors are parents. “AI is probably the best way to save time and control costs. There’s no problem with AI being involved in the initial appearance,” said Dr. Sarah Lee Davidson, Maryland, who is advocate for the Arizona bill regulating AI in previous permits. She said the anxiety has more to do with the vague pre-certification process.
Even without AI, advance permission is already a black box. Doctors often receive little or no feedback on why claims are rejected and approval metrics are constantly changing. A KFF analysis of the ACA market plan for 2023 found that the most common reason listed for denial (34%) was “no other reason listed.”
Brian Callahan, a neurologist at the University of Michigan in Ann Arbor, Michigan, recently worked for several months to acquire rituximab for her patient. The claim was initially rejected without explicit reasons or an option of appeal. In contrast, he has no problem getting his patients’ medication on veteran issues.
“I feel like private insurance is trying to deny as much as possible,” says Callaghan, the lead author of a 2024 paper to improve pre-approval using AI. Some form of cost containment is essential, but it needs to be made simple. He said having a complex process means that it means that there are fewer things to be approved because the process is challenging, not because of science or clinical adequacy.
Payers are heavily regulated in other areas, but there are few guardrails to ensure that advance permission is based on clinical data and is based on the patient’s greatest interest. And payers have a clear conflict of interest, Renato said.
It’s not the technique that makes a doctor nervous. It’s a power imbalance, Renato added. “Insurers have unlimited access to resources to deny claims and a lot of incentives to do so,” he said. “And providers often have no access to additional resources (to push back).”
Callahan agrees. The doctor has already spent a lot of time and has designated staff members who are handling advance permissions. He firmly believes that if previous approvals continue as is, AI will help solve some of these issues, but it is complicated, constantly changing, and difficult to talk to fellow doctors — AI may make the process even more difficult, he said.
Restrain your hopes
Still, doctors hope that AI can reduce the burden if advance approval becomes more transparent. This technology minimizes the time spent on documents on both sides. Also, if the initial appearance of a claim is automated, approvals can occur faster.
Space startups have already built AI-based tools to integrate bised insurance portals, speed up insurance verification before visiting, determine where prior approval is required, and automatically send electronic requests to the appropriate payer.
In the future, more sophisticated generation AI algorithms could possibly tackle negativity, Lenert said. “Even if (the AI) reflects on the rejection of a claim faster and gives an explicit reason why it becomes a service.” That would give me much more information than doctors have to continue today.
But as long as payers can say no without taking responsibility for the need to act, that’s not fair,” Renato said. With or without AI, that’s true. The industry needs new denial criteria. “If it’s not completely transparent, it’s completely wrong,” he added.
Limit point
At the state level, more laws on advance permitting continue to be in place, the NAIC reported. “It reached its breaking point,” Killerea said. Pre-authorization has shifted from debate in the health industry to public concern. “We’re already watching the state jump into the fight.”
The aim is to set up some guardrails so that the system cannot be used to delay or reject clinically appropriate care, rather than eliminating prior approval. There are also many laws specific to AI. California and Colorado have already passed legislation restricting automated decision-making systems in previous permits. Texas, Arizona, Connecticut and Indiana have posted similar bills in their work.
“There’s no problem with AI being involved in the initial look (in claim),” said Lee-Davisson of Arizona.
Zeke Silva, MD, a private practice intervention intervention radiologist in Houston, supports a similar bill in Texas. “We accept that advance approval has some role, and that algorithms have some role in patient care,” he said. “But if you are refusing to care, it should be by someone trained to practice medicine.”
Donavyn Coffey is a Kentucky-based journalist who reports on healthcare, the environment and what impacts our food. She holds her master’s degree from the Arthur L. Carter Journalism Institute at NYU in New York City and a master’s degree in molecular nutrition from Aarhus University in Aarhus, Denmark. You can see more of her work on Wired, Teen Vogue, Scientific American and more.
