Using AI to better price wildfire insurance
A startup is using machine learning and better data to more accurately assess wildfire risk for insurers.
Why it matters: Wildfires are one of the fastest-growing risks to properties, but insurers have struggled to accurately price policies — and some have abandoned the field altogether, leaving property owners unprotected.
Driving the news: The month-old Dixie Fire is now the single biggest fire in California's recorded history, and it has already destroyed more than 1,000 buildings, including nearly all of the small community of Greenville.
- Nationally, more than 4.5 million homes have been identified as being at high or extreme risk of wildfire, and on average, more than 2,500 homes per year are destroyed in wildfires.
- "In California, more than 2 million homeowners are running the risk of losing their insurance coverage because of wildfires," says Attila Toth, CEO of Zesty.ai, an Oakland-based startup that uses AI to help insurers assess climate risk.
How it works: Zesty.ai has developed a model called Z-FIRE that uses aerial imagery, data from connected devices, climate studies and machine learning to provide more accurate wildfire risk assessments.
- "In the past, risk has been explained at the regional level," Toth says. "But with the advent of Big Data, and aerial imagery and other data sources that can be analyzed with artificial intelligence, we can build models that are very precise and look at risk at the individual property level."
What to watch: This week, Zesty.ai entered into a partnership with the California FAIR Plan Association — the state's insurer of last resort — to use Z-FIRE for wildfire risk assessment.
The bottom line: Given the growing threat from wildfire in California, "insurance is a very important solution, but it's not all of the solution," says Toth.
- The state will need to consider encouraging denser housing away from wildfire risk areas, because even as Dixie rages, Californians are still ready to move to the danger zone in search of affordable housing.