Getting to grips with the misuse of AI in insurance coverage

“This year, we wanted to highlight the recurring theme of the global protection gap from a different angle – examining how the insurance industry can restore trust and deliver more societal value.”

Exploring a few of the key themes of EY’s newest ‘Global Insurance Outlook’ report, Isabelle Santenac (pictured), international insurance coverage chief at EY, emphasised the function that belief and transparency play in unlocking development. It’s a hyperlink put firmly underneath the microscope within the annual report because it examined how the insurance coverage market is being reshaped by a number of disruptive forces together with the evolution of generative AI, altering buyer behaviors and the blurring of trade strains amid the event of recent product ecosystems.

Tackling the difficulty of AI misuse

Santenac famous that the interconnectivity between these themes is grounded in the necessity to restore belief, as that is on the heart of discovering alternatives in addition to challenges amid a lot disruption. This is especially related contemplating the drive of the trade to change into extra customer-focused and improve the loyalty of shoppers, she stated, which requires clients having belief in your model and what you do.

Zeroing in on the “exponential topic” that’s synthetic intelligence, she stated she’s seeing a substantial amount of recognition throughout the trade of the alternatives and dangers AI – and significantly generative AI – presents.

“One of the key risks is how to ensure you avoid the misuse of AI,” she stated. “How do you ensure you’re using it in an ethical way and in a way that’s compliant with regulation, in particular with data privacy laws? How do you ensure you don’t have bias in the models you use? How do you ensure the data you’re using to feed your models is safe and correct? It’s a topic that’s creating a lot of challenges for the industry to tackle.”

Test circumstances or use circumstances? How insurance coverage companies are embracing AI

These challenges are usually not stopping corporations from all around the insurance coverage ecosystem engaged on ‘proof of concept’ fashions for inner processes, she stated, however there’s nonetheless a robust hesitancy to maneuver these to extra client-facing interactions, given the dangers concerned. Looking at a survey lately carried out by EY on generative AI, she famous that real-life use circumstances are nonetheless very restricted, not solely within the insurance coverage trade but in addition extra broadly.

“Everyone is talking about it, everyone is looking at it and everyone is testing some proof of concept of it,” she stated. “But no-one is really using it at scale yet which makes it difficult to predict how it will work and what risks it will bring. I think it will take a little bit of time before everyone can better understand and evaluate the potential risks because right now it’s really nascent. But it’s something that the insurance industry has to have on its radar regardless.”

Understanding the evolution of generative AI

Digging deeper into the evolution of generative AI, Santenac highlighted the pervasive nature of the know-how and the influence it’ll inevitably have on the opposite urgent themes outlined by EY’s insurance coverage outlook report for 2024. No present dialog about buyer behaviors or model fairness can afford to not discover the potential for AI to influence a model, she stated, and to look at the damaging connotations not using it appropriately or ethically may convey.

“Then on the other hand, AI can help you access more data in order to better understand your customers,” she stated. “It can help you better target what products you want to sell and which customers you should be selling them to. It can support you in getting better at customer segmentation which is absolutely critical if you want to serve your clients well. It can help inform who you should be partnering with and which ecosystems you should be part of to better access clients.”

It’s the pervasive nature of generative AI which is setting it aside from different ‘flash in the pan’ buzzwords reminiscent of Blockchain, the Internet of Things (IoT) and the Metaverse. Already AI is touching so many parts of the insurance coverage proposition, she stated, from a course of perspective, from a promoting perspective and from an information perspective. It’s turning into more and more clear that it’s a development that’s going to final, not least as a result of machine studying as an idea has already been round and in use for a very long time.

What insurance coverage corporations must be occupied with

“The difference is that generative AI is so much more powerful and opens up so many new territories, which I why I think it will last,” she stated. “But we, as an trade, want to completely perceive the dangers that come from utilizing it – bias, information privateness considerations, ethics considerations and many others. These are important dangers however we additionally want to acknowledge, from an insurance coverage trade perspective, how these can create dangers for our clients.

“For me, this presents an emerging risk – how we can propose protection around misuse of AI, around breach of data privacy and all the things that will become more significant risks with the use of generative AI? That’s a concern which is just emerging, but the industry has to reflect on that in order to fully understand the risk. For instance, experts are projecting that generative AI will increase the risk of fraud and cyber risk. So, the question for the industry is – what protection can you offer to cover these new or increasing risks?”

Insurance corporations should begin occupied with these questions now, she stated, or they run the chance of being left behind as additional developments unfold. This is particularly related on condition that some litigation has already began across the use and misuse of AI, significantly within the US. The very first thing for insurers to consider is the implications of their purchasers misusing AI and whether or not it’s implicitly or explicitly coated of their insurance coverage coverage. Insurers must be very conscious of what they’re and are usually not protecting their purchasers for, or else danger repeating what occurred in the course of the pandemic with the enterprise interruption lawsuits and payouts.

“It’s important to already know whether your current policies cover potential misuse of AI,” she stated. “And then if that’s the case, how do you want to address that? Should you ensure that your client has the right framework etc, to use AI? Or do you want to reduce the risk of this particular topic or potentially exclude the risk? I think that’s something the insurers have to think about quite quickly. And I know some are already thinking about it quite carefully.”