Western Australia’s police force has begun using live facial recognition to scan people across multiple public domains to compare their faces against watchlists.
The five-month trial uses a marked van fitted with cameras capable of identifying possible matches in real time. In its first week, the system scanned more than 130,000 people, produced 33 alerts and contributed to 19 arrests.
Police also recorded two incorrect alerts tied to the polarising technology.
Unlike traditional CCTV, which records footage for later examination, the technology converts faces into biometric data and immediately checks them against police databases.
WA Police says the system will be used during targeted and time-limited operations, with officers required to confirm every alert before taking action.
The watchlist can include people wanted on warrants, suspects in serious offences, missing people and those considered at risk of harming themselves or others.
An algorithmic alert cannot provide grounds for an arrest on its own. Officers must visually review the potential match and establish a lawful reason to intervene.
The van is clearly marked and signs warn people when facial recognition is operating. Police says biometric information belonging to people who do not generate a match is automatically deleted, and that information connected to an alert may be retained when required for an investigation, review or complaint.
Commissioner Col Blanch has described the system as a targeted policing tool rather than a general surveillance network.
“Live Facial Recognition is about protecting people, not watching people,” he said when announcing the trial.
“This technology is the way forward when it comes to supporting frontline policing.”
The published trial plan does not commit WA Police to permanent facial-recognition cameras, covert deployments or integration with the state’s broader CCTV network. The program will be assessed over five months, with its performance used to guide decisions about future use.
Police have already deployed the van at busy locations including Mirrabooka, Joondalup, Armadale and central Perth.
Blanch has also left open the possibility of using it near demonstrations where police intelligence indicates a serious risk of harm.
Any expansion beyond marked and temporary deployments would raise separate questions about oversight, watchlist criteria and the use of existing camera networks.
The facial-recognition trial is taking place alongside a broader expansion of police technology in Western Australia, but also across the globe.
WA Police is developing an automatic number plate-recognition network using fixed, temporary and mobile cameras.
Shopping centres, councils, developers and private businesses can host or help fund equipment, expanding coverage beyond government-owned locations.
The cameras can identify stolen vehicles, false number plates and vehicles associated with wanted or missing people.
WA Police has also connected hundreds of vehicles and regional stations to low-orbit satellites. The system allows officers in remote areas to retrieve number plate alerts and transmit body-camera, dashboard-camera, aircraft and security footage to Perth.
Each technology currently performs a separate function. Number plate recognition identifies vehicles, communications systems transmit footage and facial recognition compares people with police watchlists.
Together, they give police a growing ability to collect and examine information from roads, shopping centres, transport interchanges and remote areas.
WA Police has not announced plans to combine those capabilities into a single platform.
What are other countries doing?
It appears to be the inevitable way forward for governments across the developed world, with various nations implementing their own methods with rapidly advancing technology.
Britain provides one of the clearest examples of facial recognition moving from limited trials into wider police use.
Live facial recognition was initially deployed during targeted operations. The British Government has since announced 40 additional facial-recognition vans, making the technology available to every regional police force in England and Wales.
Permanent cameras have also been tested in south London, often to public distaste.
Britain’s national policing strategy envisages artificial intelligence being used to search CCTV, body-camera footage and number plate databases, analyse large quantities of information and combine separate datasets.
In Moscow, facial recognition technology has been live in various metro networks for years, using the system to identify people wanted by authorities.
In India, the city of Delhi has pursued a public-safety network involving facial-recognition analytics, person tracking and automatic number plate recognition across thousands of locations.
New York generally uses facial recognition retrospectively to analyse images connected to particular crimes rather than continuously checking people moving through the city.
However, the city also operates number plate readers, extensive camera networks, gunshot-detection sensors, police drones and the central Domain Awareness System.
Can the tech get it wrong?
Some jurisdictions have already restricted facial recognition systems due to obvious concerns about privacy.
Buenos Aires suspended its fugitive-recognition network after a court identified legal and privacy problems with its operation.
Authorities in Detroit also introduced stricter rules after facial-recognition searches contributed to wrongful arrests. Police cannot arrest somebody solely on the basis of an algorithmic match, and independent evidence must connect a suspect to an offence.
The European Union has taken a more restrictive approach under its AI Act. Real-time biometric identification by law enforcement in public spaces is generally prohibited, with limited exceptions involving missing people, imminent threats and specified serious crimes.
Deployments must generally receive prior judicial or independent approval, remain limited in time and geographic area and cannot be the sole basis for an adverse legal decision.
‘Chilling effects’: Concerns over ‘privacy-invasive’ trajectory
There has naturally been a lot of pushback from human rights campaigners across the globe as cutting-edge surveillance tech makes its way onto the streets.
Australian Privacy Commissioner Carly Kind has described facial recognition as “a highly privacy-invasive tool”, warning that it can uniquely identify people in public without their knowledge.
The Australian Human Rights Commission has gone further, arguing that “a moratorium on the use of FRT … is still needed” in decisions carrying serious legal or human-rights consequences until specific legislation is introduced.
International regulators have raised similar concerns about scale, calling for more caution from politicians and law enforcement before jumping at tools that have a smorgasbord of potential misuse cases.
The UN High Commissioner for Human Rights has warned of the technology’s “highly intrusive nature and broad impact on large numbers of people”, calling for its use in public spaces to be paused until safeguards, independent authorisation and regular human-rights audits are established.
The UN position is that it should be confined, if used at all, to serious crime or major public-safety threats.
There are also concerns about how the technology could affect protests and other lawful gatherings. UN experts have warned that scanning crowds necessarily captures the biometric information of large numbers of people who are not police targets, potentially undermining anonymity and producing a “chilling effect” on participation in public assemblies.
Former UN Special Rapporteur Clément Nyaletsossi Voule warned that facial-recognition surveillance could have “chilling effects” on society as a whole.
Former UK Biometrics and Surveillance Camera Commissioner Fraser Sampson framed the broader test more simply. Police use should be “dictated by what society agrees is acceptable, and not simply by what technology makes possible,” he said.
The Council of Europe has issued a similar warning, saying facial recognition “threatens our essential human rights, including privacy, equal treatment and non-discrimination” when deployed without strict limits.
Read related topics:Perth