The pattern hiding in plain sight
Between February and August 2025, at least five documented cases of automated parking enforcement failures emerged across New Zealand. Christchurch, Dunedin, Auckland, Lower Hutt. Each was reported as a standalone consumer gripe. None has been examined as what it actually is: a live stress test of automated decision-making in commercial settings, with results that should alarm any business deploying similar technology.
The common thread is ANPR – automatic number plate recognition – cameras that watch vehicles enter and leave carparks, then issue fines when their algorithms decide a breach has occurred. When those algorithms get it wrong, the customer’s fury lands not on the enforcement contractor but on the business whose name is on the building.
Parents, pensioners, and a system that cannot tell two visits apart
In February 2025, dozens of parents at Kindercare Wigram Skies in Christchurch received wrongful $85 fines after ANPR cameras failed to distinguish two brief drop-off visits from a single all-day park. Kindercare CEO Kelly Wendelborn estimated up to 40 parents were affected. The system had been running since August 2023, meaning errors may have been accumulating for over a year before media coverage forced a response.
A month later in Dunedin, Rod Gibbs received an $85 fine from Smart Compliance Management after two separate Kmart visits on the same day. His home security footage proved his car was in his own driveway when SCM claimed it was in the carpark. He spent half an hour trying to find a contact number to appeal because SCM had no phone number on its website.
In Auckland, Rachel Gu was fined $85 for a 100-minute grocery shop at Pak’nSave Papakura – nearly half the $209.90 she spent in store. Foodstuffs subsequently extended its time limit and added a grace period.
Meanwhile in Lower Hutt, the council’s new licence plate recognition system generated $873,000 in infringements over five months – more than triple the average parking warden’s output. A printing error sent drivers tickets with impossible timestamps showing vehicles parked from 1:32pm until 12:15pm. Nearly 200 appeals were filed.
The disability blind spot is the sharpest warning
The most damning case involves Auckland Transport’s ANPR system, launched in 2019 without consulting CCS Disability Action, the body that issues approximately 180,000 mobility permits nationwide. Because permits are issued to people rather than vehicles, plate-based recognition is structurally incapable of recognising them. CCS Disability Action reported being contacted by up to 30 permit holders who had been incorrectly fined.
AI governance commentator Dr Karaitiana Taiuru noted in August 2025 that 42,948 Auckland residents with disabled car permits are at risk, describing the failure as “a classic example of the issues that arise when stakeholders are not involved throughout the whole AI development life cycle.”
That is the profile of a discrimination complaint, not a parking dispute.
Nobody knows the error rate because nobody is counting
An OIA request to the Ministry of Justice in late 2025 sought centralised data on Disputes Tribunal claims involving private parking enforcement. The Ministry refused, not because the information was sensitive, but because it simply does not hold it. A separate OIA to NZTA found the Motor Vehicle Register was accessed over five million times in a three-month period by third parties, but NZTA cannot identify which requests were for parking enforcement purposes.
Contrast this with the state’s own automated enforcement. NZTA’s Safety Camera Network Framework mandates whole-of-system accountability, independent effectiveness measurement, and a principle that NZTA receives no direct financial benefit from fines. Average speed cameras operating under this framework achieve a 48% reduction in deaths and serious injuries. Private ANPR operators profit directly from fines, disclose no error rates, and publish no independent effectiveness data.
The retailer pays the reputational price
Here is the structural problem for any business owner reading this: Pak’nSave, Kmart, and Kindercare did not operate these enforcement systems. They contracted out. But in every case, the customer’s anger was directed at the retailer. Foodstuffs changed its policy. Kmart said it “should not have happened.” The enforcement company processed the appeal and invoiced for the next month.
As one wrongly fined Christchurch parent observed: “Vulnerable people will roll and they’ll just pay. My father would probably pay, the elderly, the people that can’t be bothered with the fight.”
The efficiency gain belongs to the operator. The reputational cost lands on the business whose name is on the door. Any company deploying automated decision-making – whether for parking, HR screening, fraud detection, or compliance monitoring – should be asking its vendor one question: what is your independently verified error rate? If the answer is silence, the next wrongful fine could carry your brand on it.
Sources
- NZ Herald: Camera glitch: Christchurch parents wrongly stung with parking fines for childcare drop-offs (2025-02-04)
- ODT: Kmart customers fume over parking glitch (2025-03-06)
- NZ Herald: Auckland mum gets $85 Pak’nSave parking fine for 100-minute shop, later waived (2025-03-28)
- NZ Herald: Mobility parking permit holder Derek Cossey was incorrectly issued fines, revealing a bigger issue (2025-03-27)
- NZ Herald: Hutt City Council’s controversial licence plate recognition parking system brings in triple the fines (2025-03-29)
- NZTA: Safety camera network framework (2024-11-01)
- FYI OIA: Statistical data regarding claims heard by the Disputes Tribunal related to private parking breach notices (2025-12-10)
- FYI OIA: Statistics on Third-Party Access to the Motor Vehicle Register (2025-12-16)
- Taiuru & Associates: Auckland Transport’s costly AI governance mistake (2025-08-05)