Marin County: PG&E Pilots AI Wildfire Reduction Program

MARIN COUNTY, CA — The Pacific Gas & Electric Company is piloting a new program in Marin County that seeks to use artificial intelligence technology to detect wildfires early.

The program was rolled out across the highest-risk fire zones in the Bay Area and across PG&E's service territory, according to a statement from the utility. The pilot couples AI technology with high-definition cameras designed to detect signs of smoke.

PG&E said it recently installed 138 HD cameras, 12 of which were located in Marin County. About 120 of those cameras were installed in the nine-county Bay Area, half of which were located in the North Bay.

The new cameras came as California continued to experience historic drought and a longer wildfire season, said PG&E Chief Risk Officer Sumeet Singh in a statement.

"Every bit of data and intelligence that comes to us could potentially save a life," he said.

The cameras offer 360-degree views and are designed to spot possible signs of smoke, then alert PG&E analysts. They determine the fire's location and whether the blaze is a wildfire, car fire or dumpster fire, said Eric Sutphin, a PG&E Wildfire Safety Operations Center supervisor, in the statement.

The technology filters out fog, dust and other phenomena that could resemble smoke.

An AI-equipped camera in Placer County spotted smoke from what would later become the River fire one minute before firefighters were dispatched to the area on August 4, the utility said.

In all, the utility installed 487 cameras since 2018 to monitor its service territory.

Months earlier, the Tubbs fire devastated Napa and Sonoma counties after private electrical equipment sparked the blaze, which scorched 36,800 acres, destroyed 5,600 buildings and killed 22 people.

PG&E said it plans to install a total of 600 smoke-spotting cameras by the end of 2022.