Crowd-Sourced Tracking of Unmarked Police Cars

As I was driving to work today, I noticed I was being tailgated. The driver seemed distracted by some electronic gizmo and wasn't paying much attention to his surroundings. I started to take a note of the car's numberplate when all of a sudden the front grill exploded with flashing blue lights, a siren started howling, and the vehicle sped off.

"Ah," I thought, "so that's what an unmarked police car looks like. I wonder if there's a way to tell if a car belongs to the police..."

I suspect that if I issued a Freedom of Information request for the licence plates of all unmarked police cars, I would be swiftly rebuffed.

That set me thinking. If I were a nefarious criminal, how hard would it be to crowd source the information?

Assume, for a moment, that I am a modern day Fagin with a veritable army of feral children to do my malicious bidding. I position the children in crime hotspots and ask them to note down the number plates of any unmarked police car which has its "blues and twos" on. The children - armed perhaps with smartphones - could faithfully record the number plate, make and model, any specific markings, and take a photo of the car.

(As an aside, there was a reasonably popular thread on the PistonHeads forum which tracked unmarked cars for a while.)

Ok, so now we have a fairly comprehensive database of unmarked police cars. What can we do with it?

Automatic Number Plate Recognition (ANPR) is the technology used by the police to spot cars and recognise their number plates. The police - or travel authority - can look up the number plate, issue fines, see if their insurance is valid, etc.

Optical Character Recognition is a complex technology, requiring massive computing power. Luckily, that power is readily available in modern smartphones.

So, as Fagin, I equip all my cars with two Android smartphones - one for the front, one for the rear. Have them continually run numberplate recognition, look up those plates in a remote database. Alert the driver if a suspected police car is near.

For more complexity, the app could check to see if the colour, make, and model match those on the database.

At the moment, one would need to have a considerable need to avoid unmarked cars and be prepared to expend a significant resource to generate the data needed in order to make this an effective counter-measure. For now, it's probably the preserve of well funded nation states.

But, buy a few autonomous drones to fly around the city - and the need for hiring expensive criminal children disappears. Computing power is only going to get cheaper, bringing this sort of activity well within the grasp of criminals everywhere.

Sousveillance is the opposite of surveillance. It's the act of watching the watchers. The above is an example of a potentially criminal use of ubiquitous computing to power a tracking infrastructure.

For people interested in civil liberties, sousveillance is a great way to keep tabs on the state - it will often use commodity hardware, crowd sourced participation, it is easily scaled, and very hard to defend against.

Technology is fundamentally morally neutral. It's our actions which determine whether a tool is used for good or evil. As technology drops in price, the infrastructure which was previously the preserve of Governments becomes available to the common man, and the common criminal.

The above could be considered an example of a movie-plot threat - it's probably not practical now, but it will be soon.

We will be living in interesting times indeed.

3 thoughts on “Crowd-Sourced Tracking of Unmarked Police Cars

  1. Jake Applebaum pointed out in his 29C3 keynote that this would work in exactly that way if you had ANPR access. More importantly, it would let you find out exactly who was being followed

    1. Possibly they would - they may well already do so. I do like the idea of fitting out all the cars with James Bond style rotating ones :-)

What do you reckon?

%d bloggers like this: