California is controlling its wildfires using AI tech

The AI system comes with caveats.
Sejal Sharma
California wildfires
California wildfires


The US state of California has recorded 4,792 and 5,234 wildfires in 2023 and 2022 respectively.

These fires burn over counties and cities, are unpredictable and climate-driven, and result in human and animal deaths, and millions and dollars worth of infrastructural damage.

The state saw its second-biggest wildfire expanding across 47,000 acres. In fact, two weeks back a mid-air collision of two choppers resulted in the deaths of three crew members, reported ABC News. They were on a mission to douse a blaze in Southern California.

The California Department of Forestry and Fire Protection, better known as Cal Fire, is on a mission to suppress 95 percent of all wildfires when they are 10 acres or less. And the department is going to use an artificial intelligence (AI) based program to detect fires across the state.

AI to the rescue

Traditionally, California’s network of over 1,000 mountaintop cameras is monitored by humans, but this is a tiring and cumbersome process. According to the officials New York Times spoke to, the AI system has been able to alert firefighters about smoke visuals around 40 percent of the time before dispatch centers received 911 emergency calls about the incident.

According to the reports by NYT and LA Times, this kind of neat AI system has so far only been installed in California. One of the drawbacks of the system is that it can only detect fires in line with the cameras. The other issue is that human interference is required to properly identify the whereabouts and veracity of the wildfire.

The project began in June this year with six Cal Fire emergency command centers having access to the AI system. The project will soon roll out to 21 centers.

The AI system is the first of its kind

The AI system is sensitive. It picks up fog, dust, and steam, and falsely identifies them as fires. The software was developed by a California-based company called DigitalPath, whose engineers have been manually looking at each fire the AI identifies. 

“You wouldn’t believe how many things look like smoke,” Ethan Higgins, a chief architect of the software, told NYT.

Operators are constantly teaching the AI machine to identify fires properly. Neal Driscoll, a geophysicist at the University of California, San Diego, and a leader of the AI project explained that the system runs its algorithm through thousands of images generated by the over 1,000 cameras deployed over 90 percent of California’s fire-prone territory.

The images are given to the AI software by a project called AlertCalifornia, a public safety program to understand natural disasters and determine short and long-term impacts on people and the environment. The project’s website is managed by UC San Diego.

“ALERTCalifornia cameras can perform 360-degree sweeps approximately every two minutes and can view as far as 60 miles on a clear day and 120 miles on a clear night,” says AlertCalifornia’s website.

The system has also proven helpful for law enforcement, as investigators can use the footage to look back to the first moments of a fire’s ignition while they work to determine its cause, reported LA Times.

Add Interesting Engineering to your Google News feed.
Add Interesting Engineering to your Google News feed.
message circleSHOW COMMENT (1)chevron
Job Board