In the summer of 2026, the Sorrento Fires scorched over 220 acres of coastal shrub in San Diego, leaving a trail of black smoke and fear. This event, along with a 20% increase in wildfires across the U.S. in 2025, highlighted the growing threat posed by hotter temperatures, prolonged drought, and expanding development. In response, scientists, firefighters, and Microsoft’s AI for Good Lab have joined forces to detect fires sooner and mitigate their impact.
Dr. Neal Driscoll, a geophysicist and founder of ALERTCalifornia at UC San Diego, shifted his focus to early fire detection after surviving the devastating Cedar Fire of 2003. His work led to the creation of ALERTWildfire in 2014, a network of fire-monitoring cameras that proved crucial during the 2019 Kincade Fire, helping evacuate over 180,000 residents safely.
In 2023, ALERTCalifornia enhanced its capabilities by integrating an AI model developed with Microsoft’s AI for Good Lab. This system analyzes footage from nearly 1,300 cameras, detecting smoke and alerting fire agencies within minutes. The AI processes over seven million images daily, providing firefighters with critical time and information to tackle blazes effectively.
The AI’s early detection has already shown significant results. In 2023, it contained the Trotter Fire in Kern County to just 52 acres, potentially preventing it from growing to nearly 4,000 acres. Zachary Wells, Deputy Chief of Kern County Fire Department, emphasized the value of the early insights provided by the system, which changes the approach to firefighting.