Western Australia launches live AI facial-recognition pilot, sparking privacy outcry
Western Australian police have begun a live facial-recognition trial that scans thousands of faces daily, prompting criticism from privacy advocates and concerns over impacts on Aboriginal communities.
In June, Western Australian police rolled out the nation’s first live facial-recognition program, deploying NEC’s Neoface m40 cameras across Perth and Fremantle to scan public faces in real time against a watchlist of roughly 4,000 suspects, missing persons and at-risk individuals. Within the first week the system generated 33 alerts, leading to 18 arrests and recorded a single false-positive identification, a rate the force touts as better than a comparable UK trial.
Critics, including academics from Melbourne Law School and the Ada Lovelace Institute, highlight known biases that cause higher error rates for women and people of colour, and note the lack of testing on First Nations faces. Aboriginal legal services argue the consultation process was superficial and that camera locations correspond to areas with significant Aboriginal populations, raising fears of over-surveillance. The Office of the Information Commissioner declined to attend a demonstration, calling the privacy impact assessment “feather-light.” Police say the trial will continue to collect demographic data to assess any disproportionate effects.
Why it matters
The trial raises questions about civil liberties, bias in AI policing and the potential over-surveillance of Indigenous communities.
In this story