Michael Terelmes, Managing Director-USA, Compre Group, moderated a panel on “The Use of AI in Workers Compensation Legacy Transactions” at the AIRROC Spring Membership Meeting. Panelists included Rooney Gleason, Chief Insurance Officer, Evolution IQ, Travis Grulkowski, Principal & Consulting Actuary, Milliman, and Justin Herring, Partner, Mayer Brown. The panelists discussed AI use in deals and the post-closing runoff process, as well as the potential regulatory hurdles facing AI usage in the legacy space. A video replay of this presentation is available on AIRROC’s On Demand platform, but here are some the top takeaways:
- AI isn’t new to the insurance industry. Some carriers on the P&C side have been using AI for years in various forms; others haven’t touched it.
- AI turns non-structured data (e.g. claim notes) into structured data, which makes it a great tool for quickly identifying problem files and assessing the value of a book of business.
- The law around AI is developing, so employ best practices when using AI in business. Consider the following: are there disclosure requirements associated with my data? Privacy concerns? Where did my data come from? Am I allowed to use my data? Can I test my model to determine if it is giving a discriminatory result? Can I correct my model if necessary?
- AI models should be constantly reviewed and retrained. They should be designed to be accurate for a specific task.
- Experienced human adjusters and large language models form a complimentary system. When using AI for claim management or review, there must be a human in the loop.
