REAL LIFE NEWS: UK POLICE USE OF FACIAL RECOGNITION ‘STAGGERINGLY INNACURATE’
by Hazed
Police in London have been using facial recognition to search crowds for people on a watch list, but a new report shows that it really doesn’t work.
The system has only correctly identified two people – and neither of those were criminals. No arrests have been made as a result of the facial recognition system.
According to information released under the Freedom of Information Act, the automated facial recognition (AFR) technology has a 98% false positive rate. That means if it identifies 100 people as on the watch list, only 2 of them will really be wanted, and 98 will be innocent.
The police claim that this figure is misleading, because once the system flags up a match, a human takes a look to check the results. But that just means that the human wastes an awful lot of time rejecting the 98 false positives.
Other police forces who submitted data on request seem to have had no better success with AFR. It seems Big Brother needs to up its game!
Meanwhile click on the second link for Dilbert’s take on facial recognition…
Source: http://www.theregister.co.uk/2018/05/15/met_police_slammed_inaccurate_facial_recognition/
http://dilbert.com/strip/2018-05-16
