Monday, May 07, 2018

The problem with facial recognition

Britain, the world's most surveilled society, is very keen on facial recognition. And you can see why: in addition to allowing people the government doesn't like to be tracked wherever they go (thanks to the UK having a camera for every 11 people), it also promises to cut the cost of policing. Instead of having to do actual police work, they can just send an arrest team whenever the cameras pop up a hit. Except there's just one problem: 90% of the "hits" are false:

A police force has defended its use of facial recognition technology after it was revealed that more than 2,000 people in Cardiff during the 2017 Champions League final were wrongly identified as potential criminals.

South Wales police began trialling the technology in June last year in an attempt to catch more criminals. The cameras scan faces in a crowd and compare them against a database of custody images.

As 170,000 people arrived in the Welsh capital for the football match between Real Madrid and Juventus, 2,470 potential matches were identified.

However, according to data on the force’s website, 92% (2,297) of those were found to be “false positives”.

So not only did it massively waste police time, it also created a real risk of innocent people being harassed or even arrested by police due to poor data. Sadly, though, this doesn't seem to have convinced them that its a bad idea.