AI in Insurance: The Good, the Bad and What Worries Regulators

Artificial intelligence has the potential to revolutionize any industry that relies heavily on data.

So it’s no wonder that the insurance industry is delving deeper and deeper into AI, to the concern and consternation of regulators everywhere.

“Insurers are all over AI,” said attorney Karen C. Yotis of the Practical Guidance Team for LexisNexis®, “because insurance impacts much of what happens in the world, and AI is infiltrating most of those areas.”

For a long while now, insurance companies have been big users of big data to calculate credit scores and develop predictive models of people’s behavior. They’re now using AI to sift through their huge datasets and unearth more insights into risks, as well as to automate countless repetitive tasks.

This growing reliance on AI, however, is making regulators nervous. Insurance regulation, by its very nature, is consumer-oriented. But AI, with its learning algorithms, operates as something like a “black box.” Nobody can be entirely sure how the AI makes its decisions.

“A challenge for regulators is understanding the complex matters they need to regulate,” Yotis said. Insurers might not even fully understand the nature or effects of their own algorithms.

This could create serious problems. Some states are already starting to act.

Insurers Using AI for Underwriting, Claims and Customer Service

Yotis said a number of states have passed laws regulating the use of AI in insurance, most notably Colorado, California, Louisiana and Connecticut. She said these laws deal primarily with credit scoring and insurance scoring, but she noted, “Legislators continue to consider new laws, and existing laws continue to evolve.”

Insurance companies are turning to AI primarily in three areas: underwriting, claims and customer service functions, although it has tremendous utility for fraud detection too.

“AI’s ability to predict fraud is unparalleled,” said Yotis, who is hosting an industry get-together on Dec. 14 at the Chicago Bar Association to discuss the growing influence of AI on the insurance sector.

Insurance companies increasingly use AI to determine if forms are filled out correctly, automate customer document requests and streamline policy drafting, all of which create “frictionless customer experiences.” Yotis said AI has a “huge” impact on reducing insurers costs and customer churn.

Handful of States Address Use of AI in Insurance in 2023

At least six states considered bills in 2023 dealing with the use of artificial intelligence in insurance. This year Colorado also became the first state to adopt a formal regulation specifically targeting the use of algorithms in insurance. That state enacted legislation in 2021 (SB21 169) restricting insurers’ use of algorithms to protect consumers from discrimination in rate-setting. 

Cons of Insurers’ Use of AI

There are also some serious potential downsides to the insurance industry’s embrace of AI.

For one, Yotis mentioned consumers’ overwhelming preference for filing claims with live humans, not automated chatbots, a partiality that even applies to younger customers who are “digital natives.” Still, insurance companies are becoming more reliant on AI for many aspects of the claims process, including at least the initial filing of a claim, which has the potential to alienate customers.

Even more concerning is the use of AI in underwriting and pricing. Insurance companies are turning to AI to analyze customers’ medical histories and lifestyle habits to assess their eligibility for coverage.

The goal, Yotis said, is for insurance companies to use AI to deepen their understanding of certain customers’ risk profiles. But she noted AI used in this way could also introduce racist or discriminatory practices, which could open insurance companies up to serious liabilities.

Class-Action Suits Over Insurers’ AI Use

Insurers have already been hit with lawsuits over their use of AI.

In 2022, State Farm was sued in the U.S. District Court for the Northern District of Illinois over claims that its AI discriminates against Black customers. The class-action suit claims State Farm's algorithms are biased against African American names.

It cited a study of 800 black and white homeowners with State Farm policies that found discrepancies in the way their claims are handled and how many delays they face.

In July, another class-action lawsuit was brought against Cigna in the U.S. District Court for the Eastern District of California over allegations that an AI algorithm it uses to screen claims was faulty and that Cigna also denied claims without having a human review them. That lawsuit was sparked by a ProPublica investigation headlined, “How Cigna Saves Millions by Having Its Doctors Reject Claims Without Reading Them.”

“These suits are hitting AI at its core,” Yotis said. “The foundational aspects of AI are being challenged.”

She said the outcome of these and other lawsuits could have far-reaching repercussions on how insurers use AI going forward.

‘Hidden Liabilities’ for Insurers

The issues involving AI in insurance are so weighty that the National Association of Insurance Commissioners (NAIC) took the unusual step of forming an AI working group to study it.

Yotis said among the concerns NAIC has identified is insurers’ heavy use of third-party data. She said NAIC is cautioning against insurers’ use of third-party data that’s been scraped or collected from websites. This is because the websites’ “terms and conditions” might not allow the use of that data, and such use could create “hidden liabilities” for insurers.

That’s just one example of the knotty issues waiting to be worked out when AI is paired with insurance. Yotis said the topic is sure to be examined in more depth as the technology evolves.

“It’s absolutely fascinating,” she said.

—By SNCJ Correspondent BRIAN JOSEPH

Please visit our webpage to connect with a State Net representative and learn how the State Net legislative and regulatory tracking solution can help you identify, track, analyze and report on relevant legislative and regulatory developments.