For years, firefighters in California have used a massive network of more than 1.000 cameras in the surrounding mountains to detect wildfires.
Operators stare at computer screens around the clock for traces of smoke. This summer, with wildfire season once again underway, California's premier fire service tries a new approach: the training of an artificial intelligence program.
The idea is to harness AI expertise to detect and prevent small fires from becoming the kinds of wildfires that have killed dozens of residents and destroyed thousands of homes in California over the past decade.
Officials involved in the pilot program say they are pleased with the early results.
About 40 percent of the time, the AI software was able to alert firefighters to the presence of smoke before fire stations received 911 calls.
"Response times have improved tremendously," said Phillip SeLegue, the chief of intelligence for the California Department of Forestry and Fire Protection, the state's primary firefighting agency, also known as Cal Fire.
In about 24 cases, Mr. SeLegue said, the AI detected fires for which the agency never received 911 calls. The fires were put out when they were still small and manageable.
The AI pilot program, which began at the end of June, covered six of its command centers Cal fire, and will be implemented in all 21 command centers from September.
Of course, the apparent success of the program is accompanied by reservations.
The system can detect fires that are only visible to cameras. At this stage, humans are still needed to make sure the AI program detects smoke correctly. Engineers at the company that created the software, DigitalPath, based in Chico, Calif., monitor the system day and night and manually check every incident the AI reports like a wildfire.
