By Darius Tahir, Lauren Sausser, KFF Well being Information
It’s the uncommon coverage query that unites Republican Gov. Ron DeSantis of Florida and the Democratic-led Maryland authorities in opposition to President Donald Trump and Gov. Gavin Newsom of California: How ought to well being insurers use AI?
Regulating synthetic intelligence, particularly its use by well being insurers, is turning into a politically divisive subject, and it’s scrambling conventional partisan strains.
Boosters, led by Trump, will not be solely pushing its integration into authorities, as in Medicare’s experiment utilizing AI in prior authorization, but additionally making an attempt to cease others from constructing curbs and guardrails. A December government order seeks to preempt most state efforts to control AI, describing “a race with adversaries for supremacy” in a brand new “technological revolution.”
“To win, United States AI firms have to be free to innovate with out cumbersome regulation,” Trump’s order mentioned. “However extreme State regulation thwarts this crucial.”
Throughout the nation, states are in revolt. A minimum of 4 — Arizona, Maryland, Nebraska, and Texas — enacted laws final 12 months reining in the usage of AI in medical insurance. Two others, Illinois and California, enacted payments the 12 months earlier than.
Legislators in Rhode Island plan to strive once more this 12 months after a invoice requiring regulators to gather knowledge on know-how use did not clear each chambers final 12 months. A invoice in North Carolina requiring insurers to not use AI as the only real foundation of a protection choice attracted important curiosity from Republican legislators final 12 months.
DeSantis, a former GOP presidential candidate, has rolled out an “AI Invoice of Rights,” whose provisions embody restrictions on its use in processing insurance coverage claims and a requirement permitting a state regulatory physique to examine algorithms.
“We’ve got a duty to make sure that new applied sciences develop in methods which can be ethical and moral, in ways in which reinforce our American values, not in ways in which erode them,” DeSantis mentioned throughout his State of the State handle in January.
Ripe for Regulation
Polling reveals Individuals are skeptical of AI. A December ballot from Fox Information discovered 63% of voters describe themselves as “very” or “extraordinarily” involved about synthetic intelligence, together with majorities throughout the political spectrum. Practically two-thirds of Democrats and simply over 3 in 5 Republicans mentioned that they had qualms about AI.
Well being insurers’ ways to carry down prices additionally hassle the general public; a January ballot from KFF discovered widespread discontent over points like prior authorization. (KFF is a well being info nonprofit that features KFF Well being Information.) Reporting from ProPublica and different information retailers in recent times has highlighted the usage of algorithms to quickly deny insurance coverage claims or prior authorization requests, apparently with little evaluate by a health care provider.
Final month, the Home Methods and Means Committee hauled in executives from Cigna, UnitedHealth Group, and different main well being insurers to deal with considerations about affordability. When pressed, the executives both denied or averted speaking about utilizing essentially the most superior know-how to reject authorization requests or toss out claims.
AI is “by no means used for a denial,” Cigna CEO David Cordani advised lawmakers. Like others within the medical insurance trade, the corporate is being sued for its strategies of denying claims, as spotlighted by ProPublica. Cigna spokesperson Justine Classes mentioned the corporate’s claims-denial course of “shouldn’t be powered by AI.”
Certainly, firms are at pains to border AI as a loyal servant. Optum, a part of well being big UnitedHealth Group, introduced Feb. 4 that it was rolling out tech-powered prior authorization, with loads of mentions of speedier approvals.
“We’re remodeling the prior authorization course of to deal with the friction it causes,” John Kontor, a senior vice chairman at Optum, mentioned in a press launch.
Nonetheless, Alex Bores, a pc scientist and New York Meeting member distinguished within the state’s legislative debate over AI, which culminated in a complete invoice governing the know-how, mentioned AI is a pure subject to control.
“So many individuals already discover the solutions that they’re getting from their insurance coverage firms to be inscrutable,” mentioned Bores, a Democrat who’s working for Congress. “Including in a layer that can’t by its nature clarify itself doesn’t seem to be it’ll be useful there.”
A minimum of some folks in medication — docs, for instance — are cheering legislators and regulators on. The American Medical Affiliation “helps state rules searching for higher accountability and transparency from industrial well being insurers that use AI and machine studying instruments to evaluate prior authorization requests,” mentioned John Whyte, the group’s CEO.
Whyte mentioned insurers already use AI and “docs nonetheless face delayed affected person care, opaque insurer choices, inconsistent authorization guidelines, and crushing administrative work.”
Insurers Push Again
With laws accepted or pending in a minimum of 9 states, it’s unclear how a lot of an impact the state legal guidelines could have, mentioned College of Minnesota regulation professor Daniel Schwarcz. States can’t regulate “self-insured” plans, that are utilized by many employers; solely the federal authorities has that energy.
However there are deeper points, Schwarcz mentioned: A lot of the state laws he’s seen would require a human to log off on any choice proposed by AI however doesn’t specify what meaning.
The legal guidelines don’t provide a transparent framework for understanding how a lot evaluate is sufficient, and over time people are inclined to turn into a little bit lazy and easily log off on any ideas by a pc, he mentioned.
Nonetheless, insurers view the spate of payments as an issue. “Broadly talking, regulatory burden is actual,” mentioned Dan Jones, senior vice chairman for federal affairs on the Alliance of Group Well being Plans, a commerce group for some nonprofit well being insurers. If insurers spend extra time working by way of a patchwork of state and federal legal guidelines, he continued, meaning “much less time that may be spent and invested into what we’re supposed to be doing, which is specializing in ensuring that sufferers are getting the suitable entry to care.”
Linda Ujifusa, a Democratic state senator in Rhode Island, mentioned insurers got here out final 12 months in opposition to the invoice she sponsored to limit AI use in protection denials. It handed in a single chamber, although not the opposite.
“There’s super opposition” to something that regulates ways corresponding to prior authorization, she mentioned, and “super opposition” to figuring out intermediaries corresponding to non-public insurers or pharmacy profit managers “as an issue.”
In a letter criticizing the invoice, AHIP, an insurer commerce group, advocated for “balanced insurance policies that promote innovation whereas defending sufferers.”
“Well being plans acknowledge that AI has the potential to drive higher well being care outcomes — enhancing affected person expertise, closing gaps in care, accelerating innovation, and decreasing administrative burden and prices to enhance the deal with affected person care,” Chris Bond, an AHIP spokesperson, advised KFF Well being Information. And, he continued, they want a “constant, nationwide strategy anchored in a complete federal AI coverage framework.”
In search of Steadiness
In California, Newsom has signed some legal guidelines regulating AI, together with one requiring well being insurers to make sure their algorithms are pretty and equitably utilized. However the Democratic governor has vetoed others with a broader strategy, corresponding to a invoice together with extra mandates about how the know-how should work and necessities to reveal its use to regulators, clinicians, and sufferers upon request.
Chris Micheli, a Sacramento-based lobbyist, mentioned the governor possible needs to make sure the state price range — constantly powered by outsize inventory market positive factors, particularly from tech firms — stays flush. That necessitates steadiness.
Newsom is making an attempt to “make sure that monetary spigot continues, and on the identical time make sure that there are some protections for California customers,” he mentioned. He added insurers imagine they’re topic to a welter of rules already.
The Trump administration appears persuaded. The president’s current government order proposed to sue and limit sure federal funding for any state that enacts what it characterised as “extreme” state regulation — with some exceptions, together with for insurance policies that defend youngsters.
That order is probably unconstitutional, mentioned Carmel Shachar, a well being coverage scholar at Harvard Legislation Faculty. The supply of preemption authority is usually Congress, she mentioned, and federal lawmakers twice took up, however in the end declined to move, a provision barring states from regulating AI.
“Primarily based on our earlier understanding of federalism and the steadiness of powers between Congress and the manager, a problem right here can be very prone to succeed,” Shachar mentioned.
Some lawmakers view Trump’s order skeptically at finest, noting the administration has been eradicating guardrails, and stopping others from erecting them, to an excessive diploma.
“There isn’t actually a query of, ought to it’s federal or ought to it’s state proper now?” Bores mentioned. “The query is, ought to it’s state or under no circumstances?”
©2026 KFF Well being Information. Distributed by Tribune Content material Company, LLC.
