Skip to main content

US woman taken down by armed police after license plate auto-reader bug

How would you feel if you were pulled over while driving and dragged out of your car at gunpoint by four policemen? And what about if the whole episode was caused by a glitch in an automatic license plate reader?

This exact situation happened to an unfortunate woman in San Francisco. And the worst part? Her car wasn't even the same make, model or colour as the wanted vehicle.

Denise Green was out driving in the city when an automated license plate reader flagged her up as driving a stolen vehicle. The police intercepted and followed her quietly for some time, even stopping behind her at a red light. Amazingly, they never bothered to compare the unfortunate victim's license plate to the one the electronic reader reported as stolen.

Then they pulled her over and forcibly removed her from her car, apparently believing her to be a hardened car thief. According to a statement, "during a portion of the time that the officers pointed their weapons at her, Green was handcuffed and secured; moreover, she weighed 250 pounds and was barely able to rise from her knees without assistance."

San Francisco police have argued that the belief that the vehicle was stolen, "in and of itself," justified the amount of force used.

The story is a parable against over-reliance on automated systems such as these.

The UK has one of the largest automatic number plate recognition systems in the world, and data kept in the system is held by police for up to two years.

Since March 2006, most motorways, main roads, town centres, London's congestion charge zone, ports and petrol station forecourts have all been covered by CCTV camera networks using automatic number plate recognition.

Existing traffic cameras in towns and cities are fast being converted to read number plates automatically as part of the new national surveillance network.

Cities across the US have also been investing in a new system known as AISight, which enables a computer system to monitor is environment, and build up a detailed profile of what can be considered "normal" behaviour. The AI can then determine what kind of behaviour is abnormal, without human pre-programming, and identify potential wrongdoers.