New Israeli AI system detects wildfires within minutes using sound sensors - interview

Wildfires pose a significant threat, as seen in the devastating fires that took nearly a year to recover from in Israel's Jerusalem Hills in 2025. As summer intensifies, the risk of new fires increases, prompting innovative solutions like the one developed by Israeli AI startup Firewave.

“

The main bottleneck with wildfires is in detection, with rescue services not being able to detect the fires before they become a major problem.

Firewave's approach tackles the critical bottleneck in wildfire response: detection. "The main bottleneck with wildfires is in detection, with rescue services not being able to detect the fires before they become a major problem," explained Dr. Jenia Yurkovsky, co-founder of Firewave. Unlike existing drone technology, Firewave utilizes sound sensors and artificial intelligence to identify fires in their nascent stages.

The system involves deploying thousands of sensors in sensitive areas. These sensors capture ambient sounds, and the AI analyzes them to differentiate between normal environmental noise and the specific acoustic signatures of a fire. "We have thousands of sounds registered that allow our system to understand not only if a fire is breaking down, but also if it's a controlled fire, a fire that was left unlit, or even a wildfire produced by the hot weather," Dr. Yurkovsky said.

“

And while technology using drones already exists, we are using sound to detect when a fire is about to happen.

This AI-powered detection is a "game-changer" because it can identify fires as small as a few square meters within minutes, a feat previously impossible due to thin smoke, low visibility in forests, and the inability of traditional systems to pick up initial sparks. "This is a game-changer because usually you detect fires when they're already starting to spread, with the first sparks not being observable by traditional detection systems," Dr. Yurkovsky stated.

“

We have thousands of sounds registered that allow our system to understand not only if a fire is breaking down, but also if it's a controlled fire, a fire that was left unlit, or even a wildfire produced by the hot weather.

The ability to detect fires so early is crucial for effective firefighting. It allows rescue services to employ targeted interventions, such as using fire extinguishers on small blazes, rather than immediately resorting to heavy machinery. The system also integrates seamlessly with autonomous firefighting technologies like drones and sprinklers, which are effective for tackling small, localized fires. Firewave is currently running pilot programs with the Kinneret Innovation Center and the Israel Nature and Parks Authority, deploying around 140 sensors across high-risk sites. The AI's adaptability was demonstrated in a California pilot, where it learned to distinguish the sound of cicadas as environmental noise, proving its ability to function effectively in diverse acoustic landscapes.

“

This is a game-changer because usually you detect fires when they're already starting to spread, with the first sparks not being observable by traditional detection systems.