How your insurance company is using AI (and why you should care)

Disclaimer: Our writers and editors used an in-house natural language generation platform to assist with portions of this article, allowing them to focus on adding information that is uniquely helpful. The article was reviewed, fact-checked and edited by our editorial staff prior to publication.

The insurance industry is increasingly adopting artificial intelligence (AI) to improve operational performance in areas like claims processing, underwriting, marketing and fraud detection. AI can enhance the speed and accuracy of these processes, but it also introduces potential issues like machine-generated errors, lack of transparency and bias. As AI becomes more integrated into the insurance sector, it’s crucial for companies to maintain transparency and for consumers to increase their understanding of AI.

How is artificial intelligence being used in the insurance industry?

Emerging trends show that AI in insurance is becoming the norm across the auto and home insurance industries, with life insurance not far behind. In surveys conducted by the National Association of Insurance Commissioners (NAIC) in 2022 and 2023, 88 percent of auto insurance companies reported that they use or plan to explore the use of AI in their operations, compared to 70 percent of home insurers and 58 percent of life insurance companies.

How are those companies actually deploying AI in their daily operations? That’s a little harder to determine, says Kathleen A. Birrane, Maryland Insurance Commissioner and member of the NAIC’s working group on artificial intelligence. “The insurance industry as a whole has been a slower adopter of AI than other industries,” Birrane says. “Part of the reason for that is because decision making by insurance companies is so heavily regulated.”

Some of AI’s more familiar use cases in insurance don’t involve regulated areas of insurance that have the potential to impact consumers. Think customer service chatbots, which companies like Progressive and Geico have been using for years. But in recent years, more and more insurers have started experimenting with AI to estimate claim amounts, approve or deny claims and underwrite insurance — all of which involve decisions traditionally made by human intelligence that are tightly regulated by state insurance laws.

While that might sound risky, it’s important to note that AI doesn’t necessarily replace human decision-making in insurance. It might be better to think of AI less as a replacement for human intelligence and more as a powerful computing tool that can support human workers in making key insurance decisions — and making them more quickly. AI’s fundamental advantage in the insurance industry is its ability to work with huge datasets that would be unmanageable for humans. Doing so allows carriers to speed up routine processes, increase accuracy and even offer innovation by freeing up human intelligence for higher-level tasks.

Accelerated underwriting: An example from the life insurance industry

Let’s look at the life insurance industry as an example: in traditional life insurance underwriting, applicants often have to provide a host of medical records in order to be approved for a policy. These could include a medical exam, blood test and detailed family history.

All of that takes time, both for you and for the insurance company — but when life insurance companies employ AI in underwriting, this process can be expedited. “What they’re trying to do is substitute all of the physical work,” says Birrane. “They’re trying to close it fast and make it easy for a generation that doesn’t want to fill out 800 forms or go in and meet with somebody and give bodily fluid samples. They just want to close the deal online.” No scale, no needle in your arm — just a profile generated by an AI program that can compare your data to a vast bank of other policyholders’ data and get you a price and a policy in record time.

How will AI affect my insurance?

It’s impossible to say exactly how AI will affect any individual’s insurance coverage in the next few years, as insurance companies continue to develop and refine AI tools. Every company’s adoption of AI and machine learning (ML) will look different, so the company that oversees your insurance matters as much as general trends in the industry. In general, however, the two trends most likely to affect consumers in the insurance industry are faster claims processing and increased accuracy in insurance pricing — both of which could come with pros and cons depending on your situation.

Faster claims processing times

AI has the potential to significantly reduce the time it takes an insurer to process a customer’s claim. According to a recent study by the Boston Consulting Group, your insurance company might use AI tools to analyze your notice of loss for missing information, summarize key information for adjusters, identify a successful negotiation strategy based on previous cases, draft communications from your adjuster and prepare reports. Home and auto insurers in the NAIC’s AI/ML surveys reported using AI tools to determine settlement amounts, pull relevant information, triage and assign cases to adjusters and evaluate loss images. In 2022, nine auto insurance companies also reported using AI to approve car insurance claims.

When carried out by human adjusters, these tasks add dozens of hours to claim processing times. Slow claim responses have historically been a major pain point for consumers, so the adoption of AI to cut down on wait times could mean an improvement in customer experiences. It could also reduce the cost of handling each claim. However, the use of AI also requires guardrails to protect against gaps that still exist within the technology.

AI software used to analyze damage is subject to errors and misinterpretation. Birrane gives the example of the use of drone imagery in processing home insurance claims. Drones are increasingly used to capture footage of homes which can then be analyzed by AI tools to approve or deny claims or determine a settlement amount. But the software is vulnerable to mistakes, such as mistaking a shadow on a roof for pre-existing damage, resulting in a claim being denied.

“People know that drone technology is being utilized generically,” Birrane says. “But do I know that they came and looked at my house?” Without transparency on the part of insurers, customers can be left in the dark and without further recourse regarding a claim denial. “I don’t have any problem with using drones,” she says. “What I want to make sure of as a regulator is that that consumer understands that a drone was used, understands what their profile was as a result of that, and has the opportunity to correct misinformation.”

More accurate pricing

As home and auto insurance rates continue to climb across the U.S., the use of AI in insurance underwriting raises both hopes and fears for consumers. Will AI’s advanced data analytics capabilities, which allow insurers to hone in on increasingly detailed individual profiles in underwriting coverage, bring prices down — or drive them up?

The answer might be a bit of both. While AI-enhanced underwriting certainly has the potential to make pricing more personalized and accurate, it’s up to insurers to choose how they want to deploy those capabilities. Some insurance companies might use AI tools to spread risk across risk pools, charging low-risk customers slightly more to offer affordable rates for high-risk customers, while others might hone in on the best risk profiles and price out less desirable customers. “It’s critically important that consumers shop, shop, shop,” Birran concludes. “Because companies have different philosophies.”

But with the promise of more accurate pricing comes the risk of discrimination, particularly the introduction of bias. As AI tools recognize patterns in customer data, they could inadvertently privilege certain classes over others if data that’s currently understood as neutral serves as a proxy for protected categories like race, ethnicity or religion.

And the fact that insurers can’t collect this type of protected demographic data will increase the challenge of identifying – and stopping – bias in AI-supported insurance underwriting. Making sure that AI tools don’t conflict with existing regulatory standards will be one of the primary focuses for the NAIC in coming years, says Birrane. She elaborates: “The question is, how do you figure out redlining, whether it’s intentional or unintentional, in an AI-driven world? Once upon a time, you could literally draw a red line and say, ‘You’re not writing anywhere there, hmm, what’s different about that area?'” In the world of AI and big data, patterns of discrimination may become harder for humans to visualize.

What is the future of AI in insurance?

As insurance companies continue to integrate AI and ML into their business practices, expect to see some of the following trends:

• Faster claim processing: Speeding up claims by automating routine parts of the process with AI could be a key strategy for companies struggling to keep customers in an increasingly pricey market.
• Mixed transparency: Some insurers will proudly put their AI credentials front and center; others might choose to take a more discreet approach to AI.
• Experiments in pricing: As AI helps insurers cut closer to accuracy with detailed customer profiles, you might see your quotes from various insurers go up or down – depending on how they’ve chosen to deploy AI.
• Personalized coverage: There’s a lot of buzz about the potential of AI to offer new types of coverage tailored to the individual needs of each customer, rather than a one-size-fits-all policy. This could look like individualized endorsements, discounts or coverage limits.
• Conversations around bias: Insurers and regulators are both working to determine what responsible AI use looks like and how to make sure companies follow best practices for governance.

One thing’s for sure: AI is here to stay in the insurance industry. But so are humans. For each of these new frontiers of AI-driven insurance, AI tools will need human intelligence to program, check and steer their operation.