Features

Artificial intelligence and the insurance industry

Insurance AI
 

by Natasha L. Rao and Nicholas T. Badalamenti and Nicole E. Wilinski   |   Michigan Bar Journal

The 21st century has seen rapid growth in technology. At the forefront of this growth is artificial intelligence (AI).

Defined as “the capability of computer systems or algorithms to imitate intelligent human behavior,”1 AI technology is becoming more commonly used by businesses to increase marketability, cost savings, and customer engagement; expand data analysis; and enhance decision-making. The insurance industry is no exception to this trend.

Now that AI is being used by the insurance industry with increased frequency, learning more about it is critical. This poses the question: What is AI, and how can we expect it to affect the insurance industry?

HOW IS THE INSURANCE INDUSTRY IMPLEMENTING AI?

One example of insurance companies taking advantage of AI technology is the use of chatbots. Originally intended to “give basic advice, check billing information, and address common inquiries and transactions,”2 chatbots have become an online interface for current and potential customers. For instance, Progressive Insurance launched a specialized chatbot allowing consumers to interact with Flo, the company’s well-known commercial spokesperson. As AI technology has improved, chatbots allow insureds and customers to interact with a virtual assistant to discuss questions and concerns spanning from applying for a new policy to filing a claim under an existing policy.3

By 2022, “more than forty insurers ha[d] incorporated chatbots into their daily business” in an effort to enhance “the customer experience by helping customers explore and purchase policies, check billing, make payments, and file claims quickly.”4

Creating Personalized Policies, Pricing, and Coverage

As AI evolved, insurance companies were quick to embrace its uses. It is now utilized in “claims processing, underwriting, [and] fraud detection [efforts].”5 One AI model that plays a particularly useful role is machine learning, a tool that uses sensors to collect data and create an information base from which it draws certain analyses. Examples of machine learning include fitness trackers, home assistants, smartphones, and smartwatches.6 Sensors on these tools collect and respond to data in large quantities. In an insurance context, data collected by these tools is analyzed and implemented to provide customers with personalized pricing, coverage options, and optimized service.7

Another example from the automotive insurance industry is the installation of sensors in insureds’ vehicles. The sensors collect drive-related data which is processed through AI machine learning and then used to assess the safety and speed of the insured and issue personalized policies and determine premiums. By tracking an insured’s “behavior behind the wheel, auto insurance companies have been able to encourage more responsible driving and try and help reduce risky behavior.”8

Smart home technology is also being used by the insurance industry.9 Sensors installed throughout insureds’ homes collect data, which is then compiled and analyzed through AI machine learning, helping providers draft policies and premiums tailored to specific risks. However, AI technology not only benefits the insurance providers, but also the homeowners. The same sensors often detect risks, such as flooding, well before any damage becomes known.10 Additionally, by “gathering and aggregating post-event data from numerous instances over time,” AI helps predict and reduce various property losses.11

In the life insurance industry, AI tools are being used to “customize coverage options and automate the underwriting process, helping to allow for flexible plans designed to fit consumer’s needs.”12

AI in Claims Handling

Insurance providers have incorporated AI into claims processing and handling, “which involve decisions traditionally made by human intelligence that are tightly regulated by state insurance laws.”13

AI tools enable providers to condense claim processing times. By using AI, insurance companies stand to save time and money and minimize errors in these processes.14 AI may allow insurers to “provide recommendations based on quick data analysis, arming agents with the right information.”15 Notably, by analyzing images, sensors, and past data, AI tools allow insurers to quickly review claims and predict potential costs.

For insurance companies, AI provides the opportunity to improve the expediency of issuing policies and handling and resolving claims.

AI’s Prominent Role in the Industry

Given the various uses of this technology, it is no surprise that AI use is being embraced by the insurance industry. Surveys conducted by the National Association of Insurance Commissioners show that in 2022 and 2023, 88% of auto insurance companies, 70% of home insurance companies, and 58% of life insurance companies were currently using or planned to use AI in their operations.16 Clearly, AI use is on the rise, and we should expect its presence to continue to expand.

POTENTIAL BENEFITS AND SETBACKS

As AI becomes more popular, it becomes more apparent that it is here to stay. Consumers and businesses alike are benefiting from this integration. However, like most new technology, AI isn’t problem-free. There are also glitches and potential downsides to its use. For anyone in or adjacent to the insurance industry, the time to consider the capabilities of AI — and its shortcomings — is now. AI’s introduction into the insurance industry has already resulted in litigation.

In 2022, a class action lawsuit was filed against State Farm in the United States District Court for the Northern District of Illinois.17 The complaint alleged that through the use of AI in claims processing and fraud detection, State Farm discriminated against Black policyholders in violation of the Fair Housing Act.18

For background, State Farm is alleged to have collected data about policyholders, including characteristics such as sex, race, gender, and education. From this data, State Farm created profiles for its policyholders that reflected their preferences, characteristics, psychological trends, and intelligence. Those profiles were then used in claims processing. Initial claims were automated through AI, which uses predictive modeling or rules-based decision making to determine whether to pay claims immediately or trigger further investigation. The AI tool learns from the data and considers how prior claims were handled to provide recommendations for handling current claims. However, as the lawsuit alleges, this process inadvertently resulted in Black claimants being subjected to more scrutiny than white claimants.

Ultimately, it was alleged that State Farm violated Section 3604(b) of the Fair Housing Act, which prohibits racial discrimination in services connected with the sale of a dwelling. The court reasoned that because some housing lenders require borrowers to have homeowners insurance, issuing homeowners insurance is a service in connection with the sale of a dwelling. This action remains pending but regardless of the outcome, it provides an example of unanticipated problems resulting from AI use.

PROS AND CONS OF AI IN THE INSURANCE INDUSTRY

On the one hand, AI use in the insurance industry provides undeniable benefits. AI can sift through substantial datasets to identify fraud-related risks. In fact, some believe that its “ability to predict fraud is unparalleled.”19 Additionally, AI has the potential to provide for a more efficient underwriting process. By automating the collection of customer data, AI reduces the time spent developing competitive and personalized insurance policies.20 Further, AI can provide more efficient claims processing. Decisions are no longer delayed through inevitable human errors but are handled through virtual assistants like chatbots that are available around the clock and can be better suited to answer customers’ questions.21

On the other hand, the increased use of AI gives rise to concerns. AI may not be bias free. Additionally, as insurance companies (and other businesses) continue to pool customer data through AI, they risk inadvertently disclosing sensitive and private information. As more data gathering occurs through AI, the incentive for hacks increases.22

HOW WILL THESE CHANGES AFFECT THE INDUSTRY?

The insurance industry is large and impacts nearly every segment of our society. Most individuals and businesses want to ensure their possessions are protected. To assist the insurance industry in providing the best service it can, we all have a role to play. Here, we call on customers, insurance providers, state legislators, and attorneys involved in insurance-related litigation to adopt procedures and best practices aimed at optimizing AI’s safe and ethical use in the industry.

All of us must be encouraged to use and test available AI tools. Only through engagement with the technology can its pitfalls and shortcomings be identified so fixes and improvements can follow.

For insurance providers and anyone using AI, establishing a proper AI risk management framework is a must. In January 2023, the National Institute of Standards and Technology released the first version of its AI Risk Management Framework,23 a great starting place for any AI user.

Proper risk management frameworks will consider the following risks:24

  • Robustness: risk of AI failing under unanticipated circumstances or cybersecurity attacks;
  • Bias: risk of AI discriminating against certain individuals on the basis of race, sex, gender, or other demographics;
  • Privacy: risk that AI discloses private or sensitive data;
  • Transparency: risk that customers are confused and uninformed throughout the claims process due to AI involvement;
  • Efficacy: risk that the AI’s intended uses will not be achieved in practice.

If a framework is established with consideration of these five risks in mind, insurance providers are more likely to achieve AI’s desired outcomes and avoid falling victim to any one of the AI-related concerns.

Legislators are also getting involved in the regulation of AI. In 2023, Colorado enacted SB21-169, entitled Restrict Insurers’ Use of External Consumer Data.25 This act prohibits an insurer from using “external consumer data and information source, algorithm, or predictive model” with regard to any insurance practice that unfairly discriminates on the basis of race, color, disability, or sex, among others.26 Under SB21-169, insurers must advise the commissioner of insurance about the uses of external consumer data and AI.27 This type of legislation may help provide guidance and regulation on AI’s use, but will also add a layer of compliance that the insurance industry and their legal advisors must address.

Lawyers will also be tasked with determining how to utilize AI in their practices as well as how to address issues related to the use of AI in claims processing from both customer and provider perspectives.

CONCLUSION

Both insurers and insureds stand to benefit from the insurance industry’s use of AI. Lawyers working in the insurance industry are wise to learn about AI and stay up to date on how AI is used by insurance companies (as well as other industries) and the extent to which that use is controlled or legislated. Much is unknown and much is to be learned about AI. But one thing is clear, AI is here to stay, and its uses are potentially endless.


ENDNOTES

1. Merriam Webster’s Dictionary https://www.merriam-webster.com/dictionary/ artificial20intelligence (all websites accessed November 19, 2024).

2. Artificial Intelligence, The National Association of Insurance Commissions https://content.naic.org/cipr-topics/artificial-intelligence.

3. Niccolo Mejia, Artificial Intelligence at Progressive-Snapshot and Flochatbot, Emerj (March 24, 2020) https://emerj.com/ai-sector-overviews/artificial-intelligence-progressive/.

4. Artificial Intelligence, supra n 2.

5. Id.

6. Ramnath Balasubramanian, Ari Libarikian, and Doug McElhaney, Insurance 2030—The Impact of IA on the future of insurance, McKinsey & Company (March 12, 2021), https://www.mckinsey.com/industries/financial-services/our-insights/insurance-2030-the-impact-of-ai-on-the-future-of-insurance.

7. Id.

8. Artificial Intelligence and the future of life insurance, Trustage Financial Group, Inc. (June 3, 2020) https://www.trustage.com/learn/life-insurance/ai-life-insurance.

9. Insurance Technology Trends: Smart Home Sensors and New Communication Tools for Consumers, Vonage America, LLC https://www.vonage.com/resources/articles/ insurance-technology-trends-smart-home-sensors-new-communication-tools-consumers/>(article updated October 27, 2022).

10. Id.

11. Id.

12. Artificial Intelligence and the future of life insurance, supra n 8.

13. Id.

14. Id.

15. Agency Forward Editorial Team, How AI is transforming the insurance industry, Nationwide Mutual Insurance Company (May 13, 2024) https://agentblog.nationwide.com/agency-management/technology/how-ai-is-transforming-the-insurance-industry/.

16. R.E. Hawley, How your insurance company is using AI (and why you should care), Bankrate (May 1, 2024) https://www.bankrate.com/insurance/car/artificial-intelligence-meets-the-insurance-industry/.

17. Huskey v State Farm Fire & Cas Co, 22 C 7014 (ND Ill, September 11, 2023).

18. Id.

19. AI in Insurance: The Good, the Bad and What Worries Regulators, AI in Insurance: The Good, the Bad and What Worries Regulators, LexisNexis (December 11, 2023) https://www.lexisnexis.com/community/insights/legal/capitol-journal/b/state-net/posts/ai-in-insurance-the-good-the-bad-and-what-worries-regulators.

20. How AI is transforming the insurance industry, supra n 15.

21. Id.

22. Mathew Tierney, Kaitlyn Ramirez, and Jeff Witmyer, Use AI in insurance — without compromising cybersecurity, Grant Thornton (January 23, 2024) https://www.grantthornton.com/insights/articles/insurance/2024/use-ai-in-insurance-without-compromising-cybersecurity.

23. Artificial Intelligence Risk Management Framework (AI RFM 1.0), National Institute of Standards and Technology (January 2023) https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf.

24. Ayesha Gulle, The Need for Risk Management in AI Systems, Holistic AI https://www.holisticai.com/blog/need-for-risk-management-in-ai (February 27, 2023).

25. See SB21-169, Colorado, (July 6, 2021) (legislation to address potential unfair discrimination by insurers in the use of third-party data).

26. Id.

27. Id.