A pioneering real-time AI facial recognition trial by Western Australia Police has sparked intense backlash from privacy watchdogs and legal advocates, who warn of racial bias, over-policing, and a severe lack of public consultation.

The Australian-first pilot, which began in June, uses sophisticated mobile cameras to scan the faces of everyday citizens in public spaces, instantly matching them against a police database. During its first week of operation in Perth and Fremantle, the system scanned more than 130,000 people.

While law enforcement authorities have hailed the initial results as a success—reporting 33 alerts and 18 arrests—civil liberties groups argue the invasive technology has been rolled out with virtually no public debate or robust regulatory guardrails.

How the Technology Works

The pilot program utilizes tech giant NEC’s "Neoface m40" platform, a complex "one-to-many" facial recognition system. High-resolution cameras scan crowds in real time, transmitting data to a nearby police van where officers monitor a live feed.

The scanned faces are cross-referenced against a watchlist of approximately 4,000 individuals, including missing persons, individuals suspected of serious offenses, and those deemed a risk to themselves or others. Police officials assert that images of individuals who do not trigger an alert are pixelated immediately and deleted from the system to protect public privacy.

Despite these assurances, experts warn that the margin for error remains a significant concern. During the first week of deployment, the algorithm falsely flagged one innocent citizen. While proponents point to this as a high accuracy rate compared to similar trials in the United Kingdom, independent researchers urge caution.

Concerns Over Algorithmic Bias

Legal and technology experts have raised alarm bells regarding the historical inaccuracies of facial recognition software, particularly concerning minority demographics.

Jake Goldenfein, an associate professor at Melbourne Law School, noted that these algorithms are frequently trained on datasets that are overwhelmingly white and male. "Historically, these systems have performed worse for people of color and for women," Dr. Goldenfein explained, adding that real-world variables like poor lighting, camera angles, and age further degrade the system's reliability.

There are also growing concerns regarding how the technology handles First Nations faces. Michael Birtwisle, an associate director of data and AI law at the Ada Lovelace Institute, warned that Australian police cannot safely rely on international safety data. He emphasized that overseas trials offer only a snapshot of specific populations and do not account for Australia's unique demographic makeup.

Aboriginal Legal Service Decries 'Tokenism'

The trial has faced fierce condemnation from the Aboriginal Legal Service of Western Australia (ALSWA). Legal Services Director Peter Collins revealed he received an email from police requesting a brief meeting to discuss the pilot just two days before it was officially launched to the public.

"This was a done deal from the get-go," Mr. Collins said, describing the sudden communication as "an exercise in tokenism."

Mr. Collins expressed deep concern that the technology would disproportionately target vulnerable communities, pointing out that the trial sites were established in areas heavily frequented by Indigenous Australians. "It can only have been by design, not chance, that facial recognition technology has been rolled out in those suburbs and places where everyone is aware that significant numbers of Aboriginal people either live or frequent," he said.

A spokesperson for the WA Police defended the program, stating that deployment locations are chosen strictly based on operational "policing need and policing objectives" rather than demographic targeting. The spokesperson added that the force is actively collating data to assess whether the trial has a disproportionate impact on any specific ethnic group.

A Lack of Independent Oversight

The regulatory rollout of the program has also drawn criticism from state watchdog agencies. The Office of the Information Commissioner (OIC) indicated it was not invited to participate in any meaningful consultation before the trial began, subsequently issuing a public warning about the dangers of deploying mass biometric surveillance without strict legislative boundaries.

John Pane, a privacy expert at Electronic Frontiers, echoed these concerns, stating that current legislation fails to keep pace with rapid technological advancements. "It grants extraordinary powers, with limited oversight, to a police agency to do incredibly invasive things," Mr. Pane warned, calling for urgent legislative reform.

As law enforcement agencies across Australia closely monitor the outcomes of the WA pilot, community advocates continue to call for an immediate halt to the trial until independent guidelines and transparent oversight measures are established.