AI Cameras Catch Fires Early
New Technology Improves Detection and
Response Times Throughout
Santa Barbara County and State
By Ella Heydenfeldt | July 17, 2025
Read more of our Fire Resiliency feature here.
The dry brush, the Santa Ana winds — the southern coast of California was designed to burn. But in a changing climate, a warming world, fires are now bigger, brighter, and much more destructive.
UC San Diego’s ALERTCalifornia program is now using artificial intelligence–powered cameras to spot wildfires early, before they grow out of control. These cameras — stationed on mountaintops across the state — watch 24/7, scanning for puffs of smoke. They don’t blink.
“That triangulation gives us a strong indication of where the fire is and allows us to start getting aircraft off the ground before units even arrive on scene,” said Captain Scott Safechuck, public information officer for the Santa Barbara County Fire Department.
Santa Barbara County now has 28 AI-equipped cameras installed across 16 sites. Each location usually hosts a pair: one camera that spins in a 360-degree patrol mode every two minutes, and another fixed on a high-risk zone selected by emergency managers.
Since the AI rollout in fall 2023, these tools have become a central part of regional fire response.
“In 2024 alone, the AI system identified 1,668 fires statewide, with 636 detections occurring before a CAD [Computer Aided Dispatch] event was created,” said CAL FIRE spokesperson Phillip SeLegue, referencing automated software that coordinates emergency responses. “These preemptive detections are critical in places like Santa Barbara, where rapid response is essential due to challenging terrain and fire-prone conditions.”
All 21 of CAL FIRE’s emergency command centers receive alerts from the AI in real time. The alerts then go to duty officers and command center staff who initiate the rollout of personnel and equipment. “This integration allows for faster assessment and resource deployment, often ahead of traditional 911 calls that are entered into our CAD notifications,” said SeLegue.
The technology is part of UCSD’s WIFIRE Lab, which started as a National Science Foundation project aimed at integrating real-time data with fire models. Now, the lab’s work fuels platforms like ALERTCalifornia, which deliver more than two billion data points per second to emergency management officials, scientists, and the public.
And while this is tech-driven, it’s not tech-dependent.
“While AI provides rapid detections, all alerts are reviewed and validated by trained personnel to ensure accuracy and avoid false alarms,” SeLegue said. “This combination enhances — but does not replace — operational situational awareness and the human element.”
“It’s definitely a coordination between human interaction, AI, and the cameras,” explained Safechuck. “It’s not something we just rely on — it’s a system that supports better decisions.”
Safechuck described a recent example that highlights the value of combining AI with human judgment. “We got a call about what someone thought was a boat on fire off the coast near Campus Point,” he said. “Our team used the cameras to check the area — and what we actually saw was just a thick marine layer. The AI didn’t flag it as smoke, but a person did. It shows how we still need that human-AI coordination to interpret what we’re seeing.”
More and more, CAL FIRE is turning to these tools as part of its long-term strategy. The agency’s goal is to suppress 95 percent of fires at 10 acres or less.
This year, CAL FIRE will deploy 100 additional cameras across the state, including forward-looking infrared (FLIR) technology that can see through smoke and the dark. Santa Barbara is part of that expansion plan.
But it stays away from people. “We blur private residences, businesses, and sensitive areas,” said Caitlin Scully, a communications manager for ALERTCalifornia. “This isn’t about watching people — it’s about protecting them.”
You must be logged in to post a comment.