WESTERN US — An artificial intelligence camera operated by Arizona Public Service detected smoke-like activity in Coconino National Forest on a March afternoon, allowing firefighters to contain the Diamond Fire before it grew beyond 7 acres (2.8 hectares). Human analysts confirmed the activity was not a cloud or dust and alerted the Arizona Department of Forestry and Fire Management and Arizona Public Service.

The utility has nearly 40 active AI smoke-detection cameras and plans to have 71 in place by summer's end. The technology notifies Arizona Public Service about 45 minutes faster on average than the first 911 call, Arizona Public Service meteorologist Cindy Kobold said.

The Arizona Department of Forestry and Fire Management has deployed seven AI cameras of its own. "Earlier detection means we can launch aircraft and personnel to it and keep those fires as small as we can," said John Truett, fire management officer for the agency.

Other utilities and agencies across the West are adopting similar tools. Xcel Energy in Colorado has installed 126 AI cameras and aims to have AI cameras in seven of the eight states it serves by year's end. ALERTCalifornia operates a network of about 1,240 AI-enabled cameras across California. The technology is mostly used in sparsely populated rural or remote high-risk areas in Arizona, California and beyond.

Pano AI, which counts Arizona Public Service among its customers, launched in 2020 and combines high-definition camera feeds, satellite data and AI monitoring. Its cameras have been deployed in Australia, Canada and 17 U.S. states including Oregon, Washington and Texas, serving forestry operations, government agencies and utilities. In the previous year, the company's technology detected 725 wildfires in the U.S. The company charges around $50,000 annually per camera, a fee that includes fire risk analysis and 24/7 intelligence center services.

The systems are not without limitations. False alarms can be costly in terms of time and attention, Patrick Roberts said, and the technology is less useful when extreme weather events like hurricane-force winds intensify and rapidly shift flames, as occurred in Los Angeles last year. In highly populated areas, people tend to spot and call in fires quickly.

"All this still requires people and decision support systems," said Roberts, a senior researcher with RAND. He said AI can also be used to identify best places to thin vegetation, conduct controlled burns and monitor air quality for signs of smoke at sensitivity about 1,000 times that of a home carbon monoxide sensor.

"AI in wildfires, it's no longer just speculative. It's really being used," Roberts said. "The future is AI everywhere, and the lines will blur between AI wildfire detection and just wildfire detection as the lines will blur in other areas of our life."