The insurance industry is increasingly leveraging artificial intelligence (AI) to enhance operational efficiency in areas such as claims processing, underwriting, marketing, and fraud detection. While AI can significantly boost the speed and accuracy of these functions, it also presents challenges, including the risk of machine-generated errors, reduced transparency, and potential biases. As AI continues to become a central component of the insurance sector, it is essential for companies to ensure transparency and for consumers to improve their understanding of AI.

How is artificial intelligence being used in the insurance industry?

Emerging trends indicate that artificial intelligence (AI) is becoming standard in the auto and home insurance sectors, with life insurance following closely behind. According to surveys by the National Association of Insurance Commissioners (NAIC) conducted in 2022 and 2023, 88% of auto insurance companies either use or plan to explore AI in their operations. This compares to 70% of home insurers and 58% of life insurance providers.

NAIC Artificial Intelligence/Machine Learning Surveys: Key Findings

  • Most auto, home, and life insurance companies in the U.S. are incorporating AI into at least one aspect of their operations.
  • 47% of home insurance companies and 18% of auto insurance companies are using AI tools for underwriting.
  • Underwriting and marketing are the primary AI applications in life insurance.
  • The top AI applications in auto and home insurance include underwriting, rating, and fraud prevention.

Determining how these companies deploy AI in their daily operations can be challenging, notes Kathleen A. Birrane, Maryland Insurance Commissioner and NAIC working group member on AI. “The insurance industry has been a slower adopter of AI compared to other sectors,” she explains, “primarily due to the heavy regulation of decision-making processes.”

While AI applications like customer service chatbots (used by companies such as Progressive and Geico) are well-known, insurers are increasingly experimenting with AI for more complex tasks like estimating claim amounts, approving or denying claims, and underwriting insurance. These functions, traditionally managed by human intelligence, are subject to stringent state insurance regulations.

Despite these advancements, AI is not a replacement for human decision-making in insurance. Instead, it serves as a powerful tool that supports human workers by processing large datasets, speeding up routine tasks, increasing accuracy, and enabling innovation. This allows human intelligence to focus on more complex tasks, enhancing overall decision-making efficiency.

Accelerated underwriting: An example from the life insurance industry

Consider the life insurance industry as an example. Traditionally, life insurance underwriting involves collecting a range of medical records from applicants, including a medical exam, blood tests, and detailed family health history. This process can be time-consuming for both the applicant and the insurance company.

However, with the integration of AI into underwriting, this process can be significantly streamlined. “The goal is to replace much of the physical paperwork,” explains Birrane. “AI aims to simplify and expedite the process for a generation that prefers to handle everything online rather than filling out extensive forms or undergoing multiple tests.” Instead of dealing with physical exams and paperwork, AI can quickly analyze a profile against a broad dataset of existing policyholders, delivering a quote and policy faster than ever before.

How will AI affect my insurance?

Predicting the exact impact of AI on individual insurance coverage in the coming years is challenging, as insurance companies are still evolving and refining their AI tools. Each insurer will adopt AI and machine learning (ML) differently, so the specifics of how AI affects your coverage will depend on the practices of your particular insurance provider. However, two key trends are expected to influence the industry: faster claims processing and improved accuracy in insurance pricing. These advancements could offer both benefits and drawbacks, depending on your individual circumstances.

Faster claims processing times

AI has the potential to greatly accelerate the claims processing time for insurers. According to a recent study by the Boston Consulting Group, insurance companies may use AI to review your loss notice for missing details, summarize essential information for adjusters, devise negotiation strategies based on past cases, draft adjuster communications, and generate reports. In NAIC’s AI/ML surveys, home and auto insurers reported using AI to determine settlement amounts, extract relevant information, triage and assign cases to adjusters, and assess loss images. By 2022, nine auto insurance companies were also using AI to approve car insurance claims.

Human adjusters typically spend many hours on these tasks, leading to slow claim responses—a common frustration for consumers. AI could significantly reduce wait times and lower the cost of handling claims, potentially improving customer experiences. However, it’s essential to implement safeguards to address the limitations of AI technology.

AI tools used for analyzing damage are not infallible. For instance, drone imagery is increasingly used to capture footage of homes for AI analysis in insurance claims. While drones provide valuable data, the AI software interpreting this data can make errors, such as misidentifying shadows as pre-existing damage, which can lead to claim denials.

“People are aware that drones are used generically,” says Birrane, “but do they know if their home was specifically inspected?” Without transparency from insurers, customers may be left unaware and unable to address inaccuracies in their claims. “I don’t object to drones being used,” Birrane adds. “What’s important is ensuring consumers know when drones are used, understand how their data is handled, and have the opportunity to correct any misinformation.”

More accurate pricing

As home and auto insurance rates continue to rise across the U.S., the role of AI in underwriting introduces both optimism and concern for consumers. AI’s sophisticated data analytics can refine underwriting processes and potentially lead to more personalized pricing, but the impact on insurance costs remains uncertain.

The effect of AI on insurance rates could be mixed. AI-enhanced underwriting may allow for more accurate pricing by better assessing individual risk profiles. However, insurers have discretion in how they apply AI. Some might use AI to balance risk across their customer base, charging slightly higher rates to low-risk individuals to subsidize rates for higher-risk customers. Conversely, others might use AI to identify and exclude less desirable risk profiles, which could drive up costs for some.

“The key is to shop around,” Birrane advises. “Different companies have different approaches and philosophies.”

However, the potential for AI to introduce bias is a significant concern. AI systems could unintentionally perpetuate discrimination if they use data that correlates with protected categories such as race, ethnicity, or religion, even though insurers cannot directly collect such demographic information.

Detecting and addressing bias in AI-driven underwriting will be a major focus for regulatory bodies like the NAIC. Birrane highlights the challenge: “In an AI-driven world, it’s not as straightforward as drawing a red line on a map and identifying areas of discrimination. The patterns of bias can be less visible and harder to pinpoint, making it crucial to ensure that AI tools comply with existing regulatory standards.”

What is the future of AI in insurance?

As insurance companies increasingly adopt AI and machine learning, several trends are emerging:

  • Faster Claim Processing: Automating routine tasks with AI could significantly speed up claims, helping companies retain customers in a market with rising prices.
  • Varied Transparency: Insurers will differ in how they disclose their use of AI; some may openly showcase their AI capabilities, while others may be more reserved.
  • Fluctuating Pricing: AI’s ability to create detailed customer profiles might cause insurance quotes to vary more widely, depending on how each company implements AI.
  • Personalized Coverage: AI promises to enable more customized insurance options, such as personalized endorsements, discounts, or coverage limits, moving beyond traditional one-size-fits-all policies.
  • Focus on Bias: Both insurers and regulators are exploring how to use AI responsibly and ensure adherence to best practices to prevent bias.

AI is becoming a permanent fixture in the insurance industry, but human oversight remains crucial. For all AI advancements, human intelligence is essential for programming, monitoring, and guiding AI systems.