Western states increasingly use AI cameras for faster wildfire detection

The Western United States is increasingly turning to artificial intelligence to spot wildfires earlier, pairing AI-monitored camera feeds with human review to verify whether a suspected smoke signal is real. The Associated Press reported that the AI systems are designed to detect visual cues resembling smoke and then alert analysts who confirm the sighting and notify agencies and utilities for faster action.

A key example cited by AP came in Arizona’s Coconino National Forest. On a March afternoon, AI flagged something resembling smoke on a camera feed, and human analysts determined it was not a cloud or dust before alerting the state’s forest service and Arizona’s largest electric utility. Firefighters raced to the scene and contained the blaze—later known as the Diamond Fire—before it grew past 7 acres (2.8 hectares), according to the report.

AP said the technology is most commonly deployed in sparsely populated, rural or remote areas where human observation may take longer, and where a blaze could go unnoticed until it expands. “It’s just the ones where we won’t get a 911 call for a long time, it is extremely helpful to have that AI always monitoring that camera,” California Department of Forestry and Fire Protection battalion chief Brent Pascua said, adding that in many cases Cal Fire responses begin before a 911 call is made.

Officials and researchers also described the urgency driving adoption. AP reported that record heat and an abysmal snowpack have raised concerns about severe wildfires, with states adding AI to their detection toolbox to protect lives and property. In Arizona, the report said Arizona Public Service has nearly 40 active AI smoke-detection cameras and plans to have 71 by summer’s end, while the state fire agency has deployed seven cameras of its own. In Colorado, AP said Xcel Energy has installed 126 cameras and aims to have cameras in seven of the eight states it serves by year’s end.

California has built a large statewide effort, AP reported. The outlet described ALERTCalifornia as a network of about 1,240 AI-enabled cameras across the Golden State, with human intervention used to keep false positives low and to improve accuracy over time. Neal Driscoll, a geology and geophysics professor at the University of California, San Diego, and founder of ALERTCalifornia, said, “The AI that’s being run on the cameras is actually beating 911 calls.”

Beyond detection, AP said commercial and research interest in AI for wildfire response has grown since companies began rolling out camera-based systems. AP reported that Pano AI, which combines high-definition camera feeds, satellite data and AI monitoring, has been deployed in Australia, Canada and 17 U.S. states, including Oregon, Washington and Texas, with customers including forestry operations, government agencies and utilities such as Arizona Public Service. The report also said Pano AI stated its technology detected 725 wildfires in the U.S. last year.

At the same time, AP reported practical limits on how the systems can be used. One obstacle is cost, with AP citing that Pano AI charges around $50,000 annually per camera, including fire risk analysis and an around-the-clock intelligence center. False alarms also present a challenge: Patrick Roberts, a senior researcher with the nonprofit research group RAND who recently finished a project on accelerating innovation in wildfire management, said false alerts can be costly in time and attention. AP also said Roberts cautioned that even accurate detections do not automatically determine the best course of action, describing how decisions about sending help, monitoring, and evacuation still require people and other decision-support systems.

Roberts said the technology’s usefulness can vary with conditions. AP reported that in more populated areas, people often spot and call in fires quickly, reducing the relative advantage of AI monitoring. The report also said the systems may be less helpful when extreme weather events, such as hurricane-force winds, rapidly shift fires—as happened in Los Angeles last year—because the human role remains crucial for adapting tactics on the ground.

In addition to helping with detection, AP said researchers are exploring broader AI applications for wildfire management, including identifying where to thin vegetation, monitoring air quality for signs of smoke, and forecasting how fires might spread and which communities could be hardest hit. AP reported that Chaowei “Phil” Yang, a professor at George Mason University in Virginia, is working with researchers from California State University of Los Angeles, the city of Los Angeles and NASA Jet Propulsion Laboratory to build a system that forecasts where a fire will burn and which neighborhoods will likely be most affected by smoke pollution. AP said the goal is to provide agencies real-time maps to inform evacuation, school and road-closure decisions, along with early air quality warnings, and that Yang said they hope to have the technology operational in three years.

Roberts said AI-enabled wildfire tools are moving beyond speculation. “AI in wildfires, it’s no longer just speculative. It’s really being used,” he said, adding that he expects adoption to continue growing.