Whether it’s underwriting, claims or customer service, it seems that artificial intelligence is top of mind for the insurance industry right now.

William Bennett, partner at law firm Saxe, Doernberger & Vita, said on this episode of The Insuring Cyber Podcast that he can see many use cases for AI in insurance, from assisting with claims handling to even drafting policy wordings. However, he cautioned that insurers need to understand how the technology works and all of the risks before investing themselves fully.

“I’m sure that the thousands, if not millions, of ways that [AI] can be utilized is catching the eyes of insurance companies in both positive and negative ways,” he said.

Bennett represents policyholders in all aspects of insurance and risk transfer and assists in the policy placement process. In this role, he helps policyholders and their brokers put together a comprehensive coverage program where all lines of coverage work together to avoid gaps. He says this work with brokers and underwriters gives him a broad practical understanding of the market and its intent.

From his perspective, generative AI in particular has been gaining more popularity within insurance regarding the use of chatbots, such as ChatGPT. Bennett said he sees an application for this technology in filling some of the talent gap in insurance, particularly on the claims side.

“It’s been pretty apparent that the industry is struggling to kind of fill its roster with enough qualified, capable people,” he said. “And a lot of what I see there is on the claims side, so I can certainly see how this could help the industry with that problem.”

However, he cautioned that while long-term benefits for the insurance industry are likely, insurers need to be careful about the many risks that can be presented when replacing human work with AI at this early stage.

“You open up a whole myriad of legal issues,” he said. “What happens when an AI tool mishandles a claim or, you know, makes some mistake in in putting together the policy? I think it’s sort of like anything else in professional services – it can be useful, but it can also be problematic if the person using it doesn’t fully understand what they’re doing themselves.”

This is likely the reason that despite some benefits, some insurers are hesitant to embrace this technology right now, he said.

“There’s a trust component there,” he said. “Until you know that it’s capable of doing the same quality work or analysis that you’re capable of doing, how can you really fully trust it and rely on it?”

As the technology evolves, one benefit that Bennett said he finds most exciting is the idea that AI chatbots in insurance could be used to draft policy wordings.

“I think the ability to mesh an understanding of all of the case law that’s out there and all of the statutes that are out there, and then also be able to read the policy form and merge those two things together, could be really incredible for writing better policy language,” he said, “if we can get to the point where [generative AI] has a good understanding of all of that governing law, which is something that’s really too big for any person to get their head around all at once at least.”

He pointed to the Insurance Services Office’s commercial general liability form – ISO CG 0001 – as one example of how AI chatbots in insurance could help with clearer policy language.

“That’s probably one of the most litigated contracts in the world, maybe the most litigated contract in the world. There are countless court opinions deciding what words and provisions in that form mean,” he said. “If someone had a good understanding of that policy form and that case law and then was asked to rewrite that form and had access to a tool that they could tell to write a provision excluding a particular risk, and tell it to account for all of the case law on that subject and write it in a way that was most true to a particular coverage intent … if that can get used properly and accessed properly, that could be really, really useful for writing clearer policy language that better accounts for all of the law out there.”

While he doesn’t expect insurance policies to be updated with exclusions regarding generative AI technology as a risk just yet, he said, he does expect that insurance regulators will likely be focusing on the use of AI chatbots in insurance in the future.

“Especially if the insurance industry starts to embrace it, I would definitely think that state insurance regulators are going to want to get ahead of that and, at a minimum, start getting some rules in place about how it could be used,” he said.

Check out the rest of the episode to hear what else Bennett had to say, and be sure to check back for new episodes of The Insuring Cyber Podcast publishing every other Wednesday along with the Insuring Cyber newsletter. Thanks for listening.