AI caught a drug trafficker just by scanning his driving patterns

Police in upstate New York recently pulled over David Zayas in the town of Scarsdale. A search of his vehicle turned up a massive stash of crack cocaine, a handgun, and $34,000 in cash. Zayas eventually pleaded guilty to drug trafficking charges.

The wild part wasn't the bust itself—it was how cops identified him in the first place. He hadn't run a red light, and his plates weren't flagged for violations. Instead, an AI algorithm simply decided his driving habits looked like those of a drug smuggler.

Scanning 16 million cars every week

According to Forbes, local authorities made the arrest using Rekor, an AI-powered roadway intelligence company that works with law enforcement agencies and private businesses across the United States.

Traditional automated license plate readers are straightforward tools. They snap photos of plates to check if a car has an expired registration or is tied to an active warrant.

Rekor takes that basic infrastructure and turns it into a massive behavioral surveillance network. In Westchester County, the system relies on 480 cameras scanning a staggering 16 million vehicles per week, logging plate numbers, makes, and models along the way.

By reverse-engineering vehicle trajectories over time, the software can reconstruct routes across entire regions and flag patterns that look statistically suspicious.

Flagged for driving like a mule

In Zayas's case, Rekor's platform analyzed his driving history over a multi-year timeline.

The algorithm spotted a clear, repetitive loop: Zayas routinely drove back and forth between Massachusetts and specific parts of upstate New York.

Crucially, the system recognized that those highways were well-documented drug trafficking corridors. It also noted that Zayas consistently made conspicuously short stays before turning around and heading back. The software concluded the pattern matched known courier behavior, giving police the tip they needed to make the stop.

Why it matters

While taking narcotics and illegal weapons off the street is an undeniable win, the way this bust happened should give everyone pause.

This isn't standard policing where an officer investigates a suspect after a crime occurs. It is bulk, dragnet data collection that subjects millions of law-abiding drivers to continuous algorithmic profiling. We think this approach is technically clever because it turns passive camera hardware into a predictive tracking grid, but the legal and privacy risks are massive.

What happens when an ordinary commuter or gig worker happens to share the travel rhythm of a narcotics courier? Rekor's tech pushes law enforcement straight into Minority Report territory, where the algorithm decides you look guilty long before an officer has reasonable suspicion.

As AI continues to supercharge public surveillance tools, the boundary between public safety and mass warrantless tracking is eroding faster than lawmakers can keep up.