Print Page | Sign In | Join
News & Press: Latest News

Is AI denying your insurance claim? It's happening way more than you think.

7 hours ago   (0 Comments)
Posted by: Diane Berg
Artificial intelligence is increasingly used by insurance companies to make decisions on claims for home repairs and medical procedures.
  • Florida is one of 22 states without specific rules or guidance addressing the use of AI in the insurance industry.
  • A failed Florida bill would have required a human to review any insurance claim denial that was initially generated by AI.
  • A new pilot program is testing the use of AI for prior authorization in traditional Medicare, which has historically had fewer restrictions than commercial plans.

You have a leaky roof, or your doctor has recommended a knee replacement to alleviate your pain. These scenarios are why you have insurance, but now the decision likely involves a machine’s determinations.

Using artificial intelligence to do your taxes, answer your questions and eliminate all varieties of tedium from your life has its appeal. However, the benefits to consumers might seem more uncertain when it's AI that is deciding what insurance will pay for your home repairs or whether your surgery will get covered.

Florida, a disaster-prone state where residents typically pay among the highest property insurance premiums in the country, is among the 22 states that have not adopted rules or guidance that specifically address AI.

At this stage, AI’s role in insurance is already spawning lawsuits and battles over its regulation, as well as a lot of anxiety about its role in the future. A class-action lawsuit against UnitedHealth Group, the nation’s largest insurer, is being closely watched. It alleges that an algorithm was used to deny nursing home care to its Medicare Advantage beneficiaries, resulting in their deaths.

Iris Smith, 80, has worked for doctors’ offices, hospitals and managed care companies, and a big reason the Delray Beach resident says she opted for the more expensive traditional Medicare plan is because the coverage doesn’t involve asking for permission, also known as preauthorization, as most commercial and Medicare Advantage plans do for a wide array of services.

Involving AI in preauthorization, as a current six-state pilot program is testing for traditional Medicare enrollees, is particularly horrifying to her.

“I don’t think a corporation … should be telling people what they can and can’t do,” said the retiree, who suffers from arthritis. "My doctors know me. I know my doctors. and when I'm in pain—which is every morning, waking up to two fists that can barely open—I need something to take care of the pain.”

What’s the role of AI in insurance?

The use of AI in insurance promises that all the rules governing insurance will be followed at a much quicker pace, industry representatives say. V7 Labs, for example, a London-based company that automates document-heavy tasks using AI agents, promises on its website to dispatch tasks that would take 30 to 60 minutes in a traditional claims process to two to three minutes.

Already, 84 percent of the nation’s health insurers are using AI for sensitive processes, like prior authorization for medical care and detecting fraud, the National Association of Insurance Commissioners found in a 2024-25 survey. NAIC is the industry’s standard-setting and support organization.

In 2023, the same organization found that 70% of property insurance companies in the U.S. were using AI or were interested in using it, and 88% of auto insurers were using or planning to use it. How well it can be relied on to interpret the nuances of human needs and evaluate individual circumstances, however, is at the forefront of political debate.

What is the AI law in Florida?

A state bill that would have required all denials to involve a human being garnered bipartisan support in the last legislative session.

“While these tools can improve speed and efficiency, no Floridian should ever have a claim denied based solely on an automated output,” said state Rep. Hillary Cassel, R-Dania Beach, who sponsored the bill that passed the House but failed to advance to the Senate.

The bill, she said, “establishes a clear and reasonable safeguard. It requires that any decision to deny a claim or portion of a claim must be made by a qualified human professional who independently reviews the facts, verifies any AI generated information and documents the basis for that decision.”

Industry representatives lined up against the bill that died. It would also seem to conflict with President Donald Trump’s executive order that no state law should regulate AI. The advancement of AI should not be hamstrung by a “patchwork of 50 different regulatory regimes,” that would put it at a competitive disadvantage with other countries’ AI development, the Dec. 11 executive order says.

Property and auto insurance regulations vary state by state, however, and it’s likely the one area that a case could be made that AI’s application to insurance should be regulated on the state level, said Gabriel Carrillo, program director for the Center for Risk Management and Insurance Education at the University of Central Florida College of Business.

“When you talk about property and casualty insurance, when you talk about life insurance, it’s almost completely regulated at the state level,” Carrillo said.

Does my insurance company use AI?

Industry representatives came before a state House committee to assure lawmakers that AI in insurance has guardrails. Although it’s a machine deciding, those decisions must comply with state rules for settling claims. They also made the case that AI’s speed and efficiency in processing claims could mean lower costs will be passed onto policyholders.

AI’s faster processing will make no difference, according to industry representatives.

“At the end of the day, who is responsible if there is a mistake in the claims-handling process? If it's an AI platform, if it's a human platform, the insurance company is always responsible, and we think that is the consistent piece here,” said Thomas Koval, a retired insurance company executive and current board member of FCCI Insurance Group, a commercial insurer out of Sarasota.

Still, an insurer’s use of AI was among the sticking points that could have meant tens of thousands of Floridians lost access to a hospital network. It was cited as health care giant Tenet Healthcare, which operates five Florida hospitals, went to the last day of its contract with Cigna, another national company, before an agreement was struck.

The contract was inked before deadline, but it was the first time AI has been a matter of contention between a health care entity as big as Tenet and an insurer as large as Cigna. Tenet said the insurance company was denying claims without human review, but Cigna publicly denied it.

Traditional Medicare piloting AI review

A pilot launched in six states this January has raised the alarm on two fronts: Traditional, fee-for-service Medicare enrollees’ procedures will be subject to more prior authorization and that authorization will be done by AI, along with human clinical review, according to federal officials.

The pilot seeks to end the waste in health care that studies have found adds up to about a quarter of the spending in U.S. health care.

 

Nearly half of the country’s 69.7 million Medicare enrollees get the traditional, fee-for-service public insurance. Traditional Medicare recipients have largely not been subject to denials that enrollees in its more recent evolution, Medicare Advantage plans, have been subject to.

KFF, a national health policy research organization, found that 53 million prior authorization requests were processed by Medicare Advantage plan insurers in 2024, resulting in 4.1 million denials. That compares with traditional Medicare. Their enrollees underwent 625,000 prior authorization reviews in 2024 and 143,705 resulted in denials—a fraction of those conducted through Medicare Advantage plans.

 

But that could be changing because of the pilot program called the Wasteful and Inappropriate Service Reduction Model (WISeR).

“The WISeR Model helps reduce clinically unsupported care by working with companies experienced in using enhanced technologies to expedite and improve the review process for a pre-selected set of services that are vulnerable to fraud, waste and abuse,” reads the press release from the Centers for Medicare and Medicare.

Evan Saltzman, a Florida State University professor in the Department of Risk Management/Insurance, Real Estate and Legal Studies, sees the pilot as part of an inevitable pull.

“It's moving traditional Medicare toward what's already done in the private sector and Medicare Advantage,” Saltzman said.

But U.S. Rep. Lois Frankel said she’s going to fight against any expansion of the effort, now being tested in six states: New Jersey, Ohio, Oklahoma, Texas, Arizona and Washington.

“We believe Medicare was based on a promise that if your doctor says you need care, if you're hurt and you need care, Medicare will be there for you, not AI,” Frankel said.

 


Anne Geggis is an Aging Well reporter for the USA TODAY NETWORK FLORIDA who focuses on physical, mental, emotional and financial well-being as we age, from Gen Y to Boomers. If you have a question you would like Anne to find answers to, send it to ageggis@usatodayco.com. You can get all of Florida’s best content directly in your inbox each weekday by signing up for the free newsletter, Florida TODAY, at https://palmbeachpost.com/newsletters