AI models, increasingly used by insurers for assessing extreme weather threats

7 November 2024 — Daniela GHETU
One-quarter of P&C insurers now deploy AI models to assess extreme weather threats, reflecting a rapid shift toward technology as climate risks increase, a research by ZestyAI, a climate and property risk analytics provider that uses machine learning to predict weather-related damage, shows.

A survey of 200 senior executives reveals that 25% of firms use AI to assess storm risk, while 18% employ it for wildfire threats. However, traditional methods remain prevalent: 54% rely on actuarial models for wildfire assessment, while 45% use stochastic models for storm risk.

The study highlights differing views on risk methods. Twenty-seven percent of executives find traditional actuarial models most accurate, 26% prefer stochastic models, and 20% trust AI models. ZestyAI CEO Attila Toth notes that extreme weather risks are driving faster AI adoption in insurance.

Executives cite peer adoption as the top factor in selecting AI models, with pricing and regulatory approval also key considerations. To mitigate climate risks, only 32% of firms prioritize non-renewals, instead focusing on inspections and deductible adjustments. Ninety percent of leaders seek more model transparency to enhance policyholder communication on risk.

As storm losses rise, convective storms are a top risk concern for 34% of insurers, with 73% of executives viewing AI as a competitive advantage.

387 views