The Zebra Newsroom

Wildfire season sparks collaboration between AI and insurers to support those in high risk states

hero image

Wildfire season has been a historically challenging market for insurance carriers, but with the seasons getting longer each year and 2020 breaking as many records as it did, insurers are seeing how climate change is directly shaping the way AI is used in insurance.

Climate-change related factors such as the prolonged drought, overall rising temperatures and buffetings have played key roles in the recent frequency and severity of wildfires and turned forests into dry fuels. In fact, the state of California has been so dry that the farmers are being told to prepare for water cutoffs to protect surrounding cities and communities from future water unavailability. A county commissioner in Oregon said that they were “looking at a drought unlike we’ve seen probably in the last 115 years.” 

Keeping insurers present in high-risk areas 

Over the past decade, insurance payouts in California have risen dramatically to reflect the increasing exposure of risk. Between 2011 and 2018, insurers paid about $4 billion annually to cover wildfire losses but the recent fire season has resulted in $26 billion in claims paid to homeowners. In addition, the shortage in builders, as well as lumber, coupled with the challenges that came with COVID-19 have slowed the rebuilding process in areas heavily affected by the wildfires. Not only does this impact insurance claims as policyholders pass on the costs of hiring builders (who have increased prices due to high demand) to their insurers, there has also been an increase in disputes regarding the actual price of reparation after wildfire damage. 

California’s insurance regulator started to discourage home building in fire-prone areas earlier this year to slow state spending on infrastructure in these zones, including looking at cutting off new construction in those regions. Commissioner Ricardo Lara endorsed proposals in June 2021 that include halting state funding for infrastructure in these areas, leaving vacant lots underdeveloped and the expansion of more stringent building codes. 

While the California Building Industry Association opposes that proposal arguing that it “places new burdens on housing development and ultimately render some new projects infeasible,” legislative committees have supported the measure. California currently has one of the most exacting building codes for areas exposed to wildfires in the most dangerous areas. Lara even mentioned that the state was hurting its existing homeowners by allowing construction to continue.“Owning a home that loses value because it’s uninsurable is really not affordable — it is a false promise that we’re making to future homeowners,” said Lara. 

Currently, there are about 4.5 million U.S. properties deemed to be at high-to-extreme risk of wildfire, with almost half (2 million) of those being in California alone. More than 1 in 12 California homes are located in high fire danger zones, and more than half-million new homes are being built in these areas. 

Some insurers are pulling out of fire-prone areas as well, which has been discouraging people from buying and selling homes in areas that are at high risk as they aren't able to access affordable insurance. This leaves homeowners without coverage, and funnels those uninsured and underinsured homeowners to get insurance through the state’s FAIR plan, which is already being stretched thin. Between 2018 and 2019 insurers dropped residential policies statewide by 31% while enrollments into the state’s FAIR plan jumped by 225%.

How AI is shaping the future of insurance and wildfire prediction 

As carriers contemplate whether or not they want to stay in the market in high-risk areas like California, tech companies have started to step up and develop tools around wildfire detection, and even act as reinsurers in these areas. SilviaTerra, a startup based in San Francisco, creates maps of the forest, outlining the ages and types of trees in each acre to prioritize where vegetation should be removed for wildfire prevention. Chooch AI, a computer-vision platform, detects the first signs of fire without the need for human involvement by using heat-sensing infrared cameras and satellite imagery. Squishy Robotics developed a robotic ball that can be dropped into a blaze and take critical temperature and wind readings that assist first responders. 

Various startups have started using advanced deep learning technology and AI to evaluate risk and provide reinsurance for carriers that cover catastrophic events like wildfires. Currently, when brand-named insurers like Allstate and State Farm experience a surge of homeowners insurance claims from catastrophic events like wildfires, reinsurers pay out about 60% of those claims. 

Farmers Insurance recently received the approval to assess wildfire risk with Zesty.ai and is expected to make standard coverage available to 30,000 homeowners. Zesty.ai uses geospatial analysis to better predict risk for individual properties and collects high-resolution images to detect roof material and vegetation that hangs over structures, as well as determine what areas were destroyed from previous incidents. Kettle uses AI, data from satellites and machine learning to predict the likelihood of a wildfire in any given square mile. By being able to predict risk better and more accurately, Kettle founders aim to keep rates affordable and offer insurance at an average 20% rate across the state. 

By integrating AI technology, pulling historical data and thinking creatively, these startups and insurtechs are revolutionizing the way we think about insurance, assessments and pricing to be more accurate, supportive and forward-thinking on behalf of policyholders. 

Other preventative measures against wildfires

State government officials in the west have been vocal about allocating more funding towards wildfire prevention as well. 

In addition to dedicating more government spending to the wildfire initiative, some high-risk states are advising residents to take preventative measures such as mapping out an evacuation plan, home hardening and preparing emergency kits.


Although the impact of 2020’s wildfire season has yet to be fully revealed, it will be interesting to see how insurance carriers, tech companies and state regulators continue to talk about how climate change is remodeling the way we think about risk, coverage, and protocol around catastrophes.

Don't forget to share!
Jasmine Kim
Jasmine KimB2B Content Manager

Jasmine is The Zebra’s newsroom content writer. With a background in journalism, she reports on breaking news, trends, mergers and acquisitions, and financial reports related to the insurance industry.