A police officer inspects the damage caused at a Pak'nSave Royal Oak supermarket, in April 2023. Photo / Hayden Woodward
A police officer inspects the damage caused at a Pak'nSave Royal Oak supermarket, in April 2023. Photo / Hayden Woodward
Opinion by Julian Benefield
OPINION
When we announced the launch of our trial of facial recognition (FR) technology in February, at 25 of our 320 stores across the North Island, we were very clear why we were doing it, and we were transparent with our approach and what it would involve.
The “why” is a retail crime crisis that’s seen our stores report over 500 breaches of trespass in just three months, and some of our team members getting stabbed, punched, kicked, bitten, spat on, sworn at, and racially abused – team members who are all someone’s partner, wife, husband, son, daughter, cousin or friend, and who are just doing their jobs.
Much of the crime is being committed by repeat offenders – around a third of all our reported retail crime, in fact – and as employers, and people who care about our customers and teams, we have a duty to do what we can to stop repeat offenders coming back to our stores.
Trialling a new and stand-alone facial recognition capability that works alongside our supermarkets’ already well-established internal security processes is about helping team members identify repeat offenders before they can do more harm.
While we appreciate the concerns raised by Consumer NZ in their recent opinion piece on our FR trial, it misses the mark in a number of areas that we wanted to address.
For starters, we’ve been very upfront about our trial before its launch and since it began in early February.
After consulting with the Office of the Privacy Commissioner, we finalised the trial’s parameters, and were deliberate in seeking as much publicity as we could so that our customers and the public are aware of it.
Consumer NZ was one of the first organisations we sent details to, and we gave the country’s most-watched daily current affairs show a behind-the-scenes demonstration of the system, so New Zealanders could see exactly how it works, why we’re trialling FR, and to get answers to their key questions.
Subsequent media coverage of our announcement reached over a million people. It was reassuring to hear Retail NZ voice their support, and to see stories in the news of shoppers also backing our move to do more to keep our stores and teams safe.
We genuinely welcomed the Office of the Privacy Commissioner’s decision to monitor our FR trial as an added layer of scrutiny, having already listened to the OPC and adjusting our trial based on the feedback received. For example, we’d appointed an independent evaluator who will monitor the trial over the coming months and help assess FR’s effectiveness in reducing crime in the trial stores. We also ensured that FR data of any shoppers who aren’t on a store’s “watchlist” would be deleted immediately.
Consumer NZ’s other concern around the idea of surveillance, in general, ignores the fact that these days the vast majority of supermarkets, petrol stations and dairies in New Zealand have security or CCTV cameras in place. Many retailers have long-recognised CCTV’s value as a crime deterrent.
The difference with FR is that instead of just passively observing people who enter a store, the FR system can automatically identify someone who has previously offended or actively assisted in offending at the store and is on the store’s watchlist. Under our trial, humans make the final decision on whether a person identified by the FR system is a previous offender or accomplice before any action is taken.
Foodstuffs announced the launch of the trial of facial recognition (FR) technology in February. Photo / 123rf
All stores taking part in the trial have large A1-sized notices (that’s eight times the size of an A4 sheet of paper) posted at their front entrances for customers to read before entering a store that’s trialling FR.
Each store’s FR system is isolated from other systems – and there is no sharing of FR data between trial stores. If you’re a shopper who is not on a store’s watchlist for having previously committed or actively assisted in retail crime, your photo and information collected will be immediately deleted from the FR system. Any claim that we are collecting and storing the biometric information of tens of thousands of citizens is misleading. No data relating to children (under the age of 18) will be stored by any of our trial stores’ individual FR systems either.
A key point in Consumer NZ’s anecdote about the friend who didn’t pay for his groceries is that it appears to have happened two years ago and is therefore unrelated to our current trial.
The current trial has been in the making for over a year – despite the exponential rise in retail crime, and literally dozens of attacks on our team members. We’ve worked hard over that period to ensure our duty to make our stores safer is rightly balanced with our obligation to respect our shoppers’ privacy.
And therein lies the answer to the question “how effective has the technology been” in reducing violence and abuse? We’re not certain yet, and that’s exactly why we are now trialling FR technology in a relatively small number of stores. We want to understand if FR can improve the safety of our team members and customers. We want to learn if the technology can help identify the individuals who think it’s okay to come into our stores and commit crime and acts of aggression towards our team members and sometimes our customers.
Our focus isn’t on petty acts of retail crime, but on the people who think it’s acceptable to intimidate, attack and injure or risk members of our team in their places of work and who have previously offended (or actively assisted in offending) at a trial store.
These repeat offenders aren’t welcome in our stores given the safety risks they bring, and we think many New Zealanders will agree. Certainly, most retailers likely do, with many watching our trial with interest.
Any suggestion that we are heavily investing in FR technology and that the trial is a forgone conclusion is incorrect.
Just 8 per cent of our stores are taking part, and we won’t make any decisions on the use of FR after our trial until we’ve consulted with our independent evaluator and the Office of the Privacy Commissioner in relation to the effectiveness of FR in reducing retail crime by repeat offenders.
Only after our independent evaluator has completed their work on our trial will Foodstuffs North Island be in a position to decide whether we continue to use, or don’t use, facial recognition, and we will continue to communicate openly and factually about our intended approach.
For more information and to watch a video about how the FR trial works click here.
