The insurance industry has been considered a ‘laggard’ to financial services in adopting new technologies such as the cloud. However, for generative AI, insurers have been quick to recognise opportunities. But what will this mean for workers and captive markets?
In a somewhat surprising turn of events, the insurance industry is emerging as a frontrunner in embracing generative AI (GenAI) technology. A 2023 Celent survey found that half of insurers had tested AI solutions, with more than 25 per cent of insurers having planned to have their own solutions in production by the end of last year. “This is a seismic shift happening in the industry, which is being catalysed by insurers making major investments into the technology,” states Richard Hartley, CEO and co-founder of insurtech Cytora.
Now, what has caused insurers to pay so much attention to the technology? Fundamentally, GenAI algorithms are built on vast quantities of data, using large language models (LLMs) to identify underlying patterns and subsequently draw conclusions or make predictions from them. This function can be applied to accelerate a wide range of insurance tasks, including underwriting.
Changes for the worker bees
Insurance is built upon data; vast amounts of risk information is unstructured and heterogeneous. These characteristics require humans to spend large quantities of time reading, scrutinising and synthesising information to understand what the risk is that they want to insure, rather than making high-level decisions based on risk. Many industry leaders have recognised the opportunity for streamlining these processes using GenAI, notes Cytora’s Hartley, aiming to free up industry workers’ time to focus on “more fulfilling activities”. For instance, Daren Rudd, UK vice president of consulting and head of the insurance business and technology consulting at CGI, has found that many speciality and commercial underwriting teams are facing too many emails to handle incoming work effectively. With AI, Rudd says, “[CGI] has been able to accelerate underwriting teams’ ability to review and understand the incoming risk and then guide the underwriters on which emails to prioritise matched to their own appetites”.
Rudd suggests that this has not only freed up time for underwriters but has also offered a rapid return on investment for “organisations that are very targetted in using this approach to handle unstructured documents”. In addition to this, industry experts have noticed a greater number of insurers using the conversational capabilities of GenAI, with specific data sets, to allow underwriting and actuarial teams to interrogate their data in a different way. These short-term implications of higher worker efficiency enabled by the technology for captives will translate into more profitable captive operations, according to Cytora’s Hartley.
When the dust settles
As the initial hype around ChatGPT starts to dissipate, the industry is taking a broader outlook at GenAI models. ChatGPT was the poster child for GenAI technology, with the platform’s seemingly limitless functions both lauded throughout 2023 and used to scaremonger, with debates around job security and the authenticity and verifiability of text and images running rife.
However, OpenAI’s ChatGPT is just one of many models built on a set of these algorithms. Many other increasingly popular applications have been built on the technology, such as Google’s Bard, for text-based AI, and DALL-E for text-to-image based AI. Some insurers have dismissed such platforms, that are available to the general public, as having limited uses for insurers due to the need to protect confidential data.
Fabian Winter, Munich Re’s chief technology officer, elaborates: “The large GenAI tools available to the general public, while promising, are of limited use to Munich Re. This is because we have highly sensitive data; we need to ensure that the knowledge generated from these data is carefully protected.” He explores this issue further in the article ‘Insurers should experiment with GenAI models now.’
Adding to Winter’s point, CGI’s Rudd says: “The ability [for the insurance industry] to deliver on the opportunities will happen if we can provide safe and secure access to the underlying data and systems that insurers already rely on.” Insurers have their work cut out to ensure efficient data protection systems are in place within their GenAI offerings, whether using in-house technology or outsourcing.
Considering the opportunity the technology presents specifically to captive insurers, Cytora’s Hartley says: “a parent company will be able to transfer more risk into the captive, and underwrite more risk”. Captives are expected to grow, as society enters a time of higher uncertainty primarily caused by climate change and geopolitical conflicts.
A previous Captive Insurance Times article, ‘CaptiveGPT’, identified the risks and shortcomings of captive insurers using ChatGPT as data overload, lack of specificity and captive insurance being so regulation-dependent.
Taking the leap
Large insurance companies have taken matters of AI into their own hands. For example, Marsh McLennan, and its management consultancy arm Oliver Wyman, have developed the company’s own GenAI tool LenAI. The tool aims to enhance employee productivity, allowing more focus on client work.
The firm says early adopters at Oliver Wyman reported that they had saved an average of eight hours per week when they used the tool. Additionally, they stated that 20 per cent less time was spent on simple, repetitive tasks and was reallocated towards more complex tasks.
These firms are one of many in the industry backing in-house models. Demonstrating this point, Cytora’s Hartley says: “We have worked with very large insurance organisations who are making fundamental investments in GenAI. They’re saying, for example, ‘We don’t want our underwriters to enter data or input information ever again’.” Eventually, underwriters could receive only rate and quote-ready risks, bypassing all manual and repetitive tasks.
In the past two decades, insurance has lagged behind other industries, such as banking, for technologies like cloud adoption. Industry leadership has recognised quickly this time around that they have a huge opportunity in GenAI.
Although, the industry mustn’t get ahead of itself – GenAI use is still in its infancy. According to CGI’s Rudd: “There is a lot more testing and investigation going on now rather than full blown use of GenAI directly in the business. A lot of work is needed in due diligence around compliance and controls to ensure the application of the technology is done with an eye on our compliance and regulatory duties.”
The due diligence workload is likely to be an even larger task for captive insurers, given the previously highlighted dependency on regulation and the levels of variation between different domiciles.
Misinformation in acceleration
Apprehensiveness around GenAI technology started to emerge and build when fake photos of Pope Francis in a puffer jacket went viral last year, prompting many to share concerns about their inability to determine whether these images are real or fake. The picture, which shows the 86-year-old head of the Catholic Church donning a mid-length white puffer jacket adorned with a silver jewelled cross, is an image generated by AI software Midjourney. First shared on discussion website Reddit, this was just one of a series of ‘deepfake’ images that left people concerned and confused.
You may be wondering how an AI-generated picture of the Pope relates to captive insurance. Well, captives can pretty much cover anything, including copyright infringement or misinformation. CGI’s Rudd spoke to one media underwriter who has been assessing potential new exposure of copyright infringement, such as using GenAI to produce images used within a film or advert.
He notes that “there has been a lot of discussion around the images these models produce to generic prompts and the similarities to golden robots and short Italian plumbers with a striking resemblance to well-known designs owned by large corporations.” Hence, there is an opportunity for captives to swoop in and fill market gaps created by these new exposures produced by AI.
The Pope picture is also emblematic of a wider concern of the increased risk of misinformation and disinformation. The WEF’s Global Risks Report 2024, produced in partnership with Marsh McLennan and Zurich Insurance Group, found that survey respondents perceive societal risks — such as misinformation and disinformation, accelerated by AI — to have increased exponentially. Close to three billion people are expected to head to the electoral polls across several economies, including Bangladesh, India, Mexico, Pakistan, the UK and the US, over the next two years. The report identifies fears that the above risk will undermine the legitimacy of newly elected governments, which could result in further civil unrest. The proliferation of the falsification of information and images will no doubt affect all industries, including captive insurance.
Job security
Among other fears around GenAI, the concern that the technology will take people’s jobs is one of the largest across all industries, and insurance is no exception.
Cytora’s Hartley assures that for the insurance industry, AI adoption will in fact have the opposite effect. “Insurance will be one of the industries where they’ll see much more opportunity to hire people for new roles,” he says.
Hartley predicts that AI will be a huge driver for economic growth. For instance, new jobs will be needed that focus on training AI platforms to perform well.
Other roles will include looking at which risks should be fully automated, which risks require human judgement and where human oversight should be applied, and how subtle changes in the risk can be identified and acted on.
CGI’s Rudd identifies a further opportunity for insurers to upskill staff and enhance their roles to provide higher value services across administration, operations, underwriting and claims with the time freed up by these new capabilities.
For captive insurers, this chance to develop their skills further could encourage more young people to join the workforce.
Excited about the prospect, Hartley says that as the proportions of their role shift away from repetitive tasks, “industry participants, particularly underwriters and claims handlers, will get more satisfaction from their roles”.
