It’s not often that insurance technology makes the paper of record. I was surprised to see an Op-Ed on AI in insurance by a member of the NYT’s editorial board this morning.
The article praised some uses of AI that contribute to customer experience like image recognition to speed claims processing. But the general tone was of fear and suspicion. AI was positioned as a scary experiment to be “left in the lab.”
The real concern is the use of advanced predictive analytics and new data sources combined with a lack of transparency. While the insurance industry has always tried to use all available data to model risk, “all available data” was pretty much limited to what distributors and prospects disclosed, plus a small number of other data sources. The information asymmetry was on the side of the customer. The insurer’s advantage was in their models, not in the richness of the data.
With the explosion of data sources and the improvements in analytics technology, the asymmetry has shifted firmly to the insurer. “All available data” is now a much more comprehensive set, and little of it is under the direct control of prospective insureds. Insurers are now capable of knowing more about prospects (and claimants) than ever before, without asking for the information directly.
While this shift in information asymmetry creates opportunities for insurers to improve risk management and pricing accuracy, it also creates both market and regulatory risk. Insurance is an industry that operates on trust. The notion that insurers are using data in a way that feels “sneaky” will be damaging to customer relationships, and may eventually result in regulatory action.
Insurers must take steps to ensure that their analytics initiatives (including AI) are not replicating unfounded and prohibited human biases. They must also take steps to communicate effectively about their use of data and analytics to the public and to regulators.
If they don’t, there’s a real risk that life and property/casualty insurers could find themselves subjected to the same restrictions as health insurers – limited to community-based rather than individual rating, and subjected to a minimum loss ratio.
Add new comment