The agency reported that the first week of the deployment resulted in 18 arrests and 33 alerts, with the algorithm producing only one false positive during that period.
The trial has faced significant criticism from privacy experts and legal advocates who warn of the infrastructure's potential for over-surveillance and demographic bias.
Associate Professor Jake Goldenfein of Melbourne Law School noted that these AI systems historically show decreased accuracy for women and people of color due to being trained on data sets of predominantly white, male faces.
Furthermore, the Aboriginal Legal Service of Western Australia characterized the police consultation as tokenistic, expressing concern that the deployment sites and the use of watchlists for outstanding warrants could disproportionately affect First Nations communities.
While the Western Australia Police maintains that images of people not on the database are pixelated and deleted almost immediately, the state’s privacy watchdog, the Office of the Information Commissioner, has criticized the lack of meaningful oversight.
The outcome of this pilot is expected to set a precedent for AI-driven law enforcement across the country, as police commissioners in other jurisdictions are monitoring the results to determine future adoption.
Moving forward, the force has committed to assessing whether the technology has a disproportionate impact on specific ethnicities or community groups.