What a Roadside Camera Can Learn
This morning, I entered my ZIP code into a website called DontGetFlocked.com, which maps reported license plate cameras. I expected to see markers in cities. I was surprised to see several out here in the country. Driving home later, I took a route that passed fewer of the marked locations.
That small change in my drive made me curious.What do these cameras actually do, and where does artificial intelligence come in?
An automated license plate reader takes a picture of a passing vehicle and records the time and place. Software reads the plate so investigators can search for it or receive an alert when a vehicle on a designated list passes a camera. The camera does not follow a car down the road like a GPS device. Each sighting is a separate record.
Flock Safety, one maker of these systems, says its software uses machine learning to identify more than a plate. It can classify a vehicle and recognize visible details such as a roof rack. Another Flock tool allows a search using an ordinary description when the plate is unknown. A witness who remembers a dark pickup with a roof rack might still provide a useful lead. Those abilities to interpret an image and search by description are specific examples of AI at work.
There are practical benefits. The City of Los Altos, California, reports that officers used a Flock sighting to locate a stolen vehicle, which another police agency recovered the same day. The city describes a separate theft investigation in which a search for a white Ford Edge with distinctive features narrowed 80 images to one matching vehicle.
What about here? A Texas Motor Vehicle Crime Prevention Authority report lists $20,000 grants in 2022 for automated license plate readers for Fayette County and, separately, the City of Schulenburg. That tells us local agencies sought this kind of investigative tool. It does not tell us who operates each camera on a public map, which company supplied it, or what rules govern today’s records.
Those rules are the part I would like to understand. Cameras record cars involved in ordinary trips as well as cars investigators may be seeking. How long are the records kept? Who may search them? Can another agency use them? Who reviews the searches? The answers may differ by agency and by the equipment it uses.
The website I visited is a starting point for questions, not an official inventory. Likewise, my taking another road is no evidence that anyone was watching me. I simply learned that technology I associated with cities may also be part of the rural landscape.
A stolen car recovered quickly is a real benefit. So is knowing how a public tool works and what limits a community has placed on it. Both deserve a place in the conversation.