The use of artificial intelligence has increased rapidly in 2023, with a third of recently surveyed companies claiming they now use generative AI (gen AI) regularly for at least one business function.  Though many consumers are concerned about the ethics and accuracy of AI,  insurance customers appear to be less resistant to it compared to when we first reported on AI and home & auto insurance.
Roughly 32% of home, auto, or renters insurance customers would trust the accuracy of AI at one or more points in their customer experience journey (application, customer service, and claims) without human oversight, and around one in seven (15%) would trust AI to handle the complete end-to-end experience, according to the 2023 Policygenius Home and Auto Insurance Technology Survey.
54% of younger insurance customers trust AI with at least some of their home or auto insurance needs
When asked if they would trust AI to process their application for insurance, answer policy questions, or assess their claim without human oversight, a majority (54%) of younger customers (age 18 to 34) said they would trust AI to execute at least one of these things accurately, compared to just 16% of customers age 55 or older.
Younger insurance customers were also the most likely to trust AI throughout the entire customer experience journey, with 26% saying they would trust an AI-driven application process, customer service, and claims assessment without a human verifying its outputs, and an additional 13% would trust AI to handle at least two of these processes without human involvement (for a total of 39%).
Meanwhile, just 6% of insurance customers age 55 or older said they would trust AI at every customer experience touchpoint, and 35% said they would not trust the accuracy of AI at all throughout the process and would want a human to handle everything, even if it led to slower application approval and delayed claim settlements.
‘Tech-savvy’ customers are more likely to trust an AI-powered insurance process
Insurers have largely moved away from paper recordkeeping and brick-and-mortar offices, allowing insurance customers to get quotes, file claims, or adjust their policy coverages without having to speak to a human. According to our survey, customers with experience doing these things online were roughly three times as likely to trust AI compared to those who had no experience with digital insurance tech.
Overview of AI use cases in home & auto insurance
Artificial intelligence has been used in the insurance industry for some time now for basic tasks, such as inputting names, addresses, and property details from an application into a spreadsheet or database. But thanks to recent advances in limited memory AI — the technology behind everything from basic chatbots to self-driving cars — AI is now capable of executing more complex tasks that we’ve normally left up to humans, such as analyzing a policy document or photographs related to a claim and making a decision about that claim based on its analysis.
Underwriting and pricing
Insurers are likely seeing the biggest benefits of AI when it comes to underwriting and issuing policies. More specifically, AI can now be used for everything from automating basic administrative tasks — like entering application data, sending reminders to customers to renew their policy, communicating with customers and loan servicers when changes have been made to a policy, and more — to quantifying and pricing the risk of an applicant by analyzing hundreds of thousands of data points. This allows underwriters to spend more of their time on applications that, for example, have already been approved by AI but require additional human verification before the company will agree to issue a policy, among other potential advantages.
On the consumer side, the increased use of AI in the application process would likely mean faster approval and more accurate and personalized rates. While one in four (25%) customers said they would trust AI to accurately process their application and receive faster approval, consumers were most likely to want a hybrid human-AI approach, with 45% saying they don’t fully trust AI to process their application accurately and would want it verified by a human, and 30% said they would not trust the accuracy of AI at all and would want a human to handle the entire process — even if it meant a slower process.
The recent boom in generative AI technology (a type of limited-memory AI) has helped insurers leverage AI in more consumer- and employee-facing ways, such as communicating with customers via chatbot, creating transcripts or course materials for insurance agents, or by conducting a sentiment analysis to ensure account specialists are prioritizing the most distressed customers and responding to them with the right level of concern.
Roughly 24% of consumers said they would trust AI to accurately answer questions about their insurance policy, 43% said they don’t fully trust AI to do this and would want the information verified by a human, and 33% said they would not trust the accuracy of AI at all and would only want a human to answer their questions.
On the claims front, limited-memory AI can be used to parse the text of a policy to see if a particular cause of loss cited in a claim is covered, or to scan photos submitted with a car insurance claim to quickly assess the wreckage and check for potential inconsistencies or fraud.
When asked about a fully AI-driven claims experience that could lead to a faster damage analysis and payout, 21% of respondents say they would trust AI to assess and process the claim accurately, 44% say they don’t fully trust the accuracy of AI and would want a human to verify the settlement, and 35% said they would not trust the accuracy of AI at all and would want a human to handle the entire claims process.
Insurance companies’ use of AI in underwriting and policy pricing, customer service, and claims could potentially improve productivity and reduce operational expenses by up to 40%, according to McKinsey.  These solutions could also lead to an overall improved user experience for home and auto insurance customers, providing them with everything from instant and hyper-personalized policy options when they submit a quote to expedited settlements when they file a claim. 
But there is also a fair amount of risk that comes with the increased use of AI in the insurance sector, namely if it will act ethically and avoid bias and discrimination when left to make decisions around applications or claims.  Furthermore, gen AI as we currently know it has a tendency to “hallucinate” from time to time and spew out inaccurate or nonsensical information, adding to concerns that it may not provide customers with correct information. 
By the end of 2023, half of insurance companies will have tested this technology, while a quarter of them are expected to have gen AI products and services already in place. 
Policygenius commissioned YouGov Surveys to poll a nationally representative sample of 2,489 adults aged 18 and older, with 2,065 having home, auto, or renters insurance. The average margin of error for responses is +/- 2%. For more details, see YouGov’s methodology. Percentages were rounded to the nearest whole number, so some totals may not add up to 100.