Police in Western Australia are now scanning faces in public and comparing them against a watchlist in real time. NZ Police is not using the technology, but with Parliament considering legislation on police image collection powers and new biometric privacy rules now in force, the question is no longer purely hypothetical.
The recent decision by Western Australia Police to trial live facial recognition technology has sparked debate about the balance between policing powers, privacy and civil liberties.
Writing in The Conversation, David Tuffley, a Griffith University applied ethics and cybersecurity senior lecturer, examined the legal and ethical issues surrounding the trial. While New Zealand's legal framework differs, many of the same questions would arise here if NZ Police was ever to consider using the technology.
Western Australia trial
Western Australia (WA) Police has deployed live facial recognition cameras in marked police vehicles around Perth. The system scans faces in public and compares them in real time with a watchlist of about 4000 people, including those wanted on outstanding warrants, registered sex offenders and missing persons, alerting nearby officers to potential matches.
WA Police says the trial is intended to improve community safety and has stressed it is not designed as mass surveillance because faces that do not match the watchlist are not retained.
Facial recognition itself is not new. Police have long used facial comparison techniques when examining CCTV footage after an offence has occurred. What makes the WA trial different is that the comparison takes place instantly, as people move through public spaces.
What would NZ law allow?
NZ Police is not using live facial recognition technology in public spaces. However, recent legal developments mean the discussion is becoming increasingly relevant.
Parliament is considering the Policing Amendment Bill, which would reaffirm and clarify Police powers to collect and keep photographs, video and audio recordings while carrying out their duties following the Supreme Court’s 2025 Tamiefuna v R decision. The Bill has cleared select committee with a majority recommending that it be passed and is now waiting for its second reading.
The proposed legislation is about collecting and retaining images. It does not expressly authorise facial recognition or artificial intelligence.
If live facial recognition were ever introduced in New Zealand, Police would still need to comply with the Privacy Act 2020 and the Biometric Processing Privacy Code 2025, which sets specific requirements around the use of biometric technologies.
The policing case
Supporters argue live facial recognition would simply automate an investigative technique Police already use – comparing CCTV images with lawfully held photographs – but in real time. Potential benefits include locating wanted offenders more quickly, identifying people breaching release conditions, finding missing people sooner and reducing the time officers spend manually comparing images during investigations.
Proponents also argue that, if appropriately limited, the technology could be targeted at specific categories of high-risk individuals rather than the public at large.
The concerns
David Tuffley identifies four principal concerns.
The first is accuracy. Even sophisticated systems can produce false matches, particularly where lighting or image quality is poor, potentially leading to innocent people being stopped or questioned.
The second is algorithmic bias. International research has found some systems perform less accurately across different demographic groups depending on the data used to train them.
Third is “automation bias” – the tendency for people to place too much confidence in computer-generated recommendations. In a policing context, it reinforces the need for any facial recognition alert to be treated as an investigative lead rather than proof of identity.
Western Australia Police has deployed live facial recognition cameras around Perth to scan faces and compares them in real time with a watchlist of about 4000 people.
Finally, Tuffley warns of “function creep”, where technology introduced for one purpose gradually expands into broader uses unless its limits are clearly defined.
What does NZ's Privacy Commissioner say?
The Office of the Privacy Commissioner has adopted a cautious rather than prohibitive approach to facial recognition technology.
Under the Biometric Processing Privacy Code 2025, agencies must demonstrate that biometric technologies are necessary, proportionate and accurate, with appropriate governance and safeguards. Similar principles informed the commissioner’s review of Foodstuffs North Island's facial recognition trial, which concluded the technology could be justified where robust safeguards existed.
What would Kiwis think?
Any future debate is unlikely to be decided by technology alone.
While the debate is often framed as privacy versus public safety, people may not see it so simply. Many people may support the use of facial recognition to locate missing persons or identify dangerous offenders, while feeling uncomfortable about routine monitoring in public places.
Questions about who could be placed on watchlists, how long information would be kept, how accurate the technology would need to be and who would oversee its use are likely to be central to public acceptance.
Privacy advocates argue that once facial recognition becomes accepted in some settings, it may become easier to expand its use into others. Supporters counter that clear rules and safeguards can prevent misuse while allowing police to benefit from new technology.
The next debate
WA’s trial demonstrates that live facial recognition is moving from theory into operational policing.
New Zealand's proposed Policing Amendment Bill would provide greater certainty around Police collecting and retaining images in public places, but it would not, on its own, authorise live facial recognition.
For the Police Association, the immediate issue is still restoring operational certainty following Tamiefuna. Whether Aotearoa eventually embraces the next technological step – and under what safeguards – is a separate debate that has barely begun.
Ultimately, the issue is one of trust. If New Zealand ever considers using live facial recognition, the debate is likely to focus not just on what the technology can do, but on who controls it, what safeguards are in place and whether the public is comfortable with its use.