Western Australia Police announced that during a five-month trial of real-time facial recognition technology, which concluded on Saturday, 1,380,882 faces were scanned. The trial resulted in 122 arrests and nine false alerts, with the technology used in 100 deployments across 43 regional and metropolitan locations.
In addition to the 122 arrests, alerts led to contact with 146 registered sex offenders and the location of 10 individuals for welfare checks. Deployments included three operations at the Perth Royal Show, where 223,749 faces were scanned, resulting in 15 alerts, three arrests and seven interactions with registered sex offenders.
A police spokesperson confirmed that the system generated nine false alerts out of a total of 1.38 million scans, equating to an error rate of 0.00065 per cent. Western Australia Police Commissioner Col Blanch confirmed that officers contacted each individual in cases of false alerts, stating that no one was arrested and no one complained about the false alerts. Blanch acknowledged that the technology is not perfect and that authorities knew this when entering the trial.
Chief Executive Officer of Civil Liberties Australia, Bill Rowlings, said that Western Australia Police is seriously whitewashing the picture by claiming a false alert rate of 0.00065 per cent. Rowlings suggested that the rate should be calculated by comparing the number of arrests that resulted in convictions. He stated that the true error rate will be known when the 122 arrested individuals appear in court, and that if half of the 122 arrested are not convicted, the error rate will be 50 per cent. Rowlings also said that police carried out nearly 1.4 million scans without any reasonable suspicion.
The state’s independent privacy watchdog, the Office of the Information Commissioner WA (OIC WA), announced that it was not involved in the implementation of the trial and warned of collective harms to privacy. Blanch stated on Sunday that the trial was focused on the responsible use of modern technology to identify individuals who pose a risk, while allowing the majority of people to carry out their daily activities without any impact.
According to the trial guidelines, cameras mounted on or near marked police vehicles scanned faces in specific public areas, with faces compared against an alert list compiled from police databases. Police announced that images of individuals not on the list were pixelated in real time and immediately deleted without storage. The trial previously sparked negative reactions due to concerns about privacy and algorithmic bias, with experts claiming that the technology could disproportionately affect marginalised groups who are already overrepresented on police surveillance lists. Facial recognition algorithms historically have higher false match rates for people with darker skin tones.
Surveillance vans have now been removed from public deployment while police conduct an internal evaluation, which includes reviewing safeguards and whether all conditions were met.
