Artificial intelligence could transform property insurance by improving risk assessments and underwriting efficiency, but major concerns exist over the lack of a meaningful regulatory apparatus and the technology's ability to reflect human biases, experts say.

As climate change-driven natural disasters have shown the limits of traditional risk assessment techniques that rely on historical data, a number of companies have pushed regulators to allow underwriters to use predictive models that could incorporate AI. Meanwhile, regulators and consumer advocates say there must be guardrails to ensure that carriers use transparent AI models that don't result in discriminatory outcomes.

While AI experts caution that the technology may be sensationalized to a certain extent, its ability to process massive amounts of information and inputs makes it a clear candidate to revamp an industry that depends on actuarial science. From claims processing to underwriting, to the chatbots consumers interact with on websites, some proponents say AI will deliver both for carriers and customers, while other experts are more circumspect about the prospect of carriers using AI to the benefit of consumers.

"Many insurance companies — most — are publicly owned companies," said William G. Passannante, an Anderson Kill policyholder attorney. "So the customer often isn't the beneficiary of those efficiencies, it's the shareholder, and that's frankly the problem with the dual-headed nature of the responsibilities of insurance company management."

Passannante told Law360 that as property insurance companies start integrating AI more meaningfully into their operations, he would expect to see more litigation brought by people claiming an AI system disadvantaged them. Still, Passannante said he wouldn't expect such litigation immediately, but rather on a more midterm horizon.

"There are already reports of algorithms making errors with regard to facial recognition at a higher rate with certain racial groups than others," he said. "How do you audit that?"

...

To read the full article, click here.