Several states are considering regulating how insurers can use AI in prior authorization.
In January, a California law took effect that prohibits payers from making coverage decisions based solely on artificial intelligence algorithms. The law requires any denial, delay or modification of care to be reviewed by a physician or other qualified healthcare provider.
Other states are considering similar actions. In 2025, lawmakers in Illinois, Georgia, Texas, Rhode Island, Minnesota and Florida have introduced laws that would restrict the use of AI in prior authorization decisions. None of the bills have passed their state legislatures as of March 26.
Insurers’ use of AI in prior authorization has come under scrutiny in recent years. UnitedHealth Group is facing a lawsuit alleging the company used AI algorithms to wrongfully deny postacute care for Medicare Advantage patients.
In 2023, an Optum spokesperson told Becker’s that naviHealth, the AI tool at the center of the lawsuit, was not used to make coverage decisions.
“The tool is used as a guide to help us inform providers, families and other caregivers about what sort of assistance and care the patient may need both in the facility and after returning home,” the spokesperson said. “Coverage decisions are based on CMS coverage criteria and the terms of the member’s plan.”
Humana is facing a similar lawsuit alleging it wrongfully used naviHealth algorithms. A spokesperson told Becker’s the company uses “‘human in the loop’ decision-making whenever AI is utilized.
Some federal lawmakers have urged greater oversight of AI in prior authorization as well. In 2024, CMS issued guidance to Medicare Advantage plans on the use of AI in coverage decisions. The agency ruled insurers can use algorithms to support coverage decisions, but they must ensure that an algorithm or an AI-based tool is compliant with the agency’s coverage decision requirements.
Mehmet Oz, MD, President Donald Trump’s pick for CMS administrator, told senators AI could “play a vital role” in speeding up prior authorizations, but the overall number of procedures that require a prior authorization should be reduced.
“I would argue that to use AI wisely, we would make a decision which is we’re only going to pre-authorize 1,000 procedures,” Dr. Oz told senators during a March 14 hearing.
Dr. Oz also said he believes CMS should use AI programs to oversee insurers use of AI.
“If we see that there’s something being done, for example, inappropriate use of AI or inappropriate denial of services with AI, we should be using AI within the agency to identify that early enough that we can prevent it and we should do it real time, not six months down the road,” he said.
