Police cameras in Western Australia scanned 1,380,882 faces during a five-month trial of real-time facial recognition that ended on Oct. 31, the Guardian reported, citing WA police; the Guardian also reported that the trial resulted in nine incorrect alerts.[1]

The fight now is over how to count those errors: a police spokesperson put the error rate at 0.00065% by measuring the nine false alerts against every scan, while Bill Rowlings, chief executive of Civil Liberties Australia, said police were “severely gilding the lily” and that the true rate would only be known once cases reach court.[1] The surveillance vans have been removed from public deployment while the trial is evaluated, the Guardian reported, after the trial scanned the faces of more than a million people.[1]

Under the trial guidelines, cameras mounted on or near marked police vehicles scanned faces in designated public areas and matched them against an alert list drawn from police databases, the Guardian reported.[1] Police said images of people not on the list were pixelated in real time and deleted immediately without being stored.[1] Officers review every alert before deciding whether to approach someone, ID Tech reported.[2] The system is NEC's Neoface m40, according to Biometric Update.[3]

All of those false matches misidentified people with darker skin tones, the outlet reported.[3] Experts cited by the Guardian say facial recognition algorithms have historically had higher false-match rates for people with darker skin tones.[1]

Blanch called the false alerts “a very, very, very low number” and acknowledged the technology was “not perfect”, the Guardian reported.[1]

The force's internal evaluation includes “reviewing the safeguards and whether all requirements were met”, the Guardian reported.[1] An Equality Impact Assessment from May 2026 commits to analysing false alerts and demographic patterns, according to Biometric Update.[3]