← CyberAdX Video · myprivacy.blog
License Plate Readers: 1.4 Billion Errors a Month, Real Consequences
A breakdown of the accuracy claims behind automated license plate readers (ALPR) and how a 7% error rate translates into billions of misidentifications monthly. Intended for privacy advocates, policymakers, and anyone concerned about surveillance tech and civil liberties.
Transcript
Is this a 2 or a 7? Is this a 0 or an 0? To a license plate reader camera, it can be hard to tell. Regardless of a mistake, the camera will still take a picture, still try to identify the car, and still report the data. In its marketing materials, the leading ALPR company claims their cameras accurately capture 93 out of 100 license plates that pass by. That means that 7 out of 100 license plates are incorrectly captured. And if, as claimed, they scan 20 billion plates per month, that means they have 1 .4 billion incorrect captures every month. Combining these numbers with human error, it's no surprise that people are being wrongly pulled over, held at gunpoint, or even put behind bars for almost a month. If the system misidentifies your license plate or incorrectly reports your car's make and model, you could trigger a hot list for a crime you didn't commit. With all these cameras, law enforcement is watching you, suspect or not. It is past time to put constitutional guardrails on their use.