As brazen acts of cargo theft pop up with increasing frequency and freight fraudsters lie through their teeth with stolen credentials and identities to get right through guard houses and pilfer loads, it feels like even in 2025, technology has not kept up with the dark side of the trucking world.
But some would argue the next big twist in tech is just around the corner with the dawn of widespread so-called "artificial intelligence."
What if AI could monitor security feeds with an electric eye and sweep fences with sophisticated sensors? What if it could make sure the right truck came to pick up the right load? Birdseye Security Solutions, a remote security company, has trained its modular system to do exactly that.
"In general we do the work security guards used to do," said Milan Luketic, chief technology officer at Birdseye. The tech is evolving fast, but the company hopes its both a security windfall for shippers and receiver and, for truckers, a time saver.
[Related: AI for owner-ops? 'Hey Bubba' AI wants to deal with brokers for you]
In about 200 truck yards and warehouses across the U.S., Birdseye does some combination of watching security cameras, scanning and identifying trucks approaching the gate and monitoring fences with sensors, even sometimes deploying countermeasures against unwanted visitors.
"In general cargo theft is becoming a bigger and bigger problem," said Luketic. "iPhones are like bars of gold now."
Birdseye's "typical customer" is any operation with a truck yard and typically a warehouse on-site. These include some smaller indepedent operations as well as big businesses like FedEx, DHL and Norfolk Southern. The system combines physical security, like "surveillance along the fence line" that "sometimes uses radar or lidar to detect any objects or people trying to enter the yard" as well as plenty of cameras, sometimes as many as 120 at any one site, Luketic said. "It's like having 120 eyeballs" watching the camera.
From the perspective of a driver showing up at one of the AI-monitored truck yards, the experience would vary, as "the bigger yards are more automated," he added. Broadly speaking, though, Birdseye works with a yard management system to learn which trucks are allowed in ahead of time.
"The AI will scan the truck for the truck numbers and match that against the authorization list, and if it's expecting truck it will lift the barrier arm," said Luketic. "In other cases we require a facial recognition or facial match with an ID," with a human to "actually do that verification" as regulations restrict filming people's faces.
Birdseye thus offers a range of security-automation options to its customers. Typically, "there needs to be a human" present in a gate house anyway, Luketic added, given how many different forms fraud can take. "One of the main ways thieves try to get in the gate is by impersonating an electrician or construction worker, or even medical staff like an ambulance driver," he said. "You need a human to engage those kind of people."
Birdseye works with two or three cameras scanning the truck, sometimes looking for DOT numbers, sometimes a plate number. "Sometimes a shadow from the mirror can block the number" from the camera, "so we do have an array of cameras to deal with that," he said. Interactions move faster with automated scanning of your credentials than a security guard, typically. Sometimes a driver will have to talk to a speaker box, but otherwise, the experience shouldn't be that unique to Birdseye overall. Luketic hopes someday the system even evolves to have "drones chaperoning drivers to their spot."
Importantly, the entire time the driver interacts with the guard shack, the system's AI with its cameras and sensors monitors the perimeter.
[Related: Owner-op jailed after unwitting participation in cargo theft: Cautionary tale, mitigating risk]
What about the edge cases, though? How does Birdseye avoid calling the cops on an innocent bystander who just happened to lean on the wrong fence?
"To be frank, right now most of the time our detection alerts will go out if there’s a pedestrian close to a fence where they shouldn’t be," he said. "If a pedestrian is walking along grassy area with no sidewalk, that will heighten the sensors for us" and ultimately refer it to a human.
But soon, Luketic said, "state of the art technology is actually crossing that chasm." AI will soon "start detecting a behavior's intent," but for now, Birdseye leans towards creating "false positives" and letting real people make the final call. The system does employ tried-and-true low-tech countermeasures, however, like a strobe light or loudspeakers if someone walking gets too close.
