Essay | American Fire Departments Are Short on People. AI Might Be the Answer.

Myke Cole, a volunteer firefighter in New York’s Hudson Valley, is a former intelligence and law enforcement officer and the author, most recently, of “Steel Lobsters: Crown, Commonwealth, and the Last Knights in England.”

One morning this past July, I was awakened at 4 a.m. by my pager. I turned on my car’s blue emergency lights and raced to the fire house. Within minutes, four fellow volunteer firefighters and I arrived at a multistory house in New York’s Hudson Valley to see flames blasting through the roof.

The first thing to do on scene was assign a firefighter to walk 360 degrees around the burning structure.

This is critical because the people who call 911 are often panicked and notoriously unreliable. They might see fire blowing out the front of the house but be totally unaware of a victim hanging out the window on the opposite side.

Still, there were only a few of us and we were facing 20-foot flames. When seconds count and only three or so people show up to a call, a lieutenant can be forgiven for assigning crew members to “fire attack” or a primary search, rather than peeling off a desperately needed body to check out the scene.

It wasn’t the first time I wished for an AI-powered drone.

In New York, the state I live in, volunteer firefighter numbers fell by a third over the past two decades, according to estimates by the Firefighters Association of the State of New York (FASNY). That’s a big drop when you consider the fact that outside of New York City, more than 80% to 85% of fire departments are staffed by volunteers, according to the same organization.

Many of these volunteers, like me, are over 50: men who are aging, even old, taking on brutal physical work.

The reasons for this decline are not well understood. Whether it constitutes a crisis is also hotly debated. A spokesperson for FASNY said that there is not a public-safety crisis but that there are certain departments throughout the state that are struggling, mostly with funding. The organization also noted that there is no central database that tracks volunteer firefighters so all participation numbers are estimates.

What is clear is that the fire service isn’t likely to see an uptick in volunteers anytime soon.

Meanwhile, the country is in the midst of a technological epoch. AI and the robotics and drones it powers are increasingly seen as threats to swaths of the workforce because they eliminate the need for humans to do certain jobs. In the volunteer fire service, ironically, that is probably exactly what we need.

To be clear, no machine could put out fires on its own: At the fire I faced in July, it was my job to feed hose into the house, keeping it from getting snagged on corners and moving loops up the stairs to the second floor to ensure the nozzleman had enough slack to reach the fire. It’s complicated work.

What’s more, fully charged with water, a 1.75-inch attack line can weigh up to 75 pounds per 50-foot length. My muscles burned and my back screamed as I raced up and down the stairs, dragging the hose around door frames.

AI couldn’t do that. But it could be a “force multiplier,” as we used to say in the military.

For example: What if a 911 dispatcher or an AI agent could launch AI-driven drones to the scene of a fire to “360” the structure, feeding back video or still images from every approach?

This could happen while the fire company was getting geared up. Firefighters would arrive on scene with a ready picture of the structure from all angles, saving critical time and keeping critical manpower on fire attack.

This is somewhat related to how some reconnaissance drones are being used in modern warfare: Drones deployed in the war Russia started with Ukraine are capable of autonomous environmental perception, navigation, mapping and thermal imaging of structures.

Drones have already been useful to firefighting even without AI. The Colorado Springs Fire Department started using drones for aerial surveillance in 2023. The U.S. Forest Service uses drones to start controlled burns on inaccessible terrain.

Reconnaissance drones in civilian service would need safeguards, as they could be ripe for exploitation by law enforcement, in much the same manner as the Flock camera network. But where augmented surveillance tools might be troubling from a civil liberties perspective, they have unarguable safety benefits.

AI drones could, for example, improve pre-incident plans, the documents prepared by firefighters to prevent fires. To make them, firefighters go to homes and businesses in their district to interview residents, understand how they live and map the premises.

These plans can be pulled up on a computer tablet while firefighters are driving to a fire, giving greater visibility into the context of an emergency situation. But the manual, time-intensive process of creating the plans often simply doesn’t get done, or is limited to the largest homes and businesses in a district, which are usually covered as training exercises.

Between public records of parcel maps, Google street view and photos that could be uploaded by consenting residents and business owners, AI could make short work of this task, potentially providing a database of pre-incident plans for every structure in a fire district, even making them shareable with neighboring fire companies.

There was no pre-incident plan for the property we encountered in the July fire. It had already been burning for some time when we showed up on scene; we stopped the flames but couldn’t save the structure.

I later learned that a few officers had arrived before our fire engine and were able to do the critical 360 reconnaissance. But I didn’t know that when our rig sounded three long horn blasts as pieces of the roof came down—a signal to evacuate and move to defensive firefighting.

Other volunteer firefighters arrived to help with the brutal work of “pulling ceilings” and opening walls to extinguish hot spots and ensure the fire would not rekindle. Despite the lack of hands, I joined two other men riding a utility truck back to the station to shower. The denuded crew was annoyed, but understood. We all know you can get fired for being late to your day job.