Published on the 10/12/2024 | Written by Heather Wright

Trial shows FRT reduces harm and justifies use…

The use of facial recognition technology during a Foodstuffs North Island trial has been deemed justified by a report which found ‘strong quantitative evidence’ it reduced the number of ‘serious harm’ events, including assault, by around 16 percent.

The reduction is the equivalent of around 100 serious harm events, including assaults, abuse and other aggressive or disorderly conduct.

“Retailers have been watching the trial with great interest and a number are investigating FRT for their own operations.”

The six month trial was instigated after the Privacy Commissioner asked Foodstuffs to show evidence that the use of facial recognition technology (FRT) was justified and made a ‘practical and statistically significant difference’ relative to other less intrusive options.

Foodstuffs North Island says it had 1,742 positive alerts across the six month trial which ran in 25 New World and Pak’nSave stores, and ‘got it right’ more than 98 percent of the time, allowing it to identify repeat offenders and intervene early before they caused trouble.

In nine instances people were approached and found to be the wrong person, including two which resulted in the person being asked to leave, something Foodstuffs says was attributable to failure in the human process involved.

In April, a Rotorua woman was mistakenly identified as a thief and kicked out of a New World store and has taken her case to the Human Rights Review Tribunal, claiming the technology is discriminatory. Accuracy of recognition on minority populations has been a key concern for the technology, and Consumer NZ says Foodstuffs’ trial did not gather information on the ethnicity of those misidentified.

Adam Barker, director of analytics firm Scarlatti, which did the final evaluation report for the Foodstuffs trial, says around half of the harm reduction can be attributed to actual interventions by supermarket staff, who approached repeat offenders before they could do more harm. The remainder came from the deterrent effect, stopping offenders from returning to the stores.

He says the results justify the use of FRT, which has proved to be a somewhat controversial offering.

In the UK some civial rights organisations are campaigning against the use of the live FRT technology by retailers. It has been widely used in the UK and other countries to help identify shoplifters. A biometric program, dubbed Project Pegasus and backed by some of the UK’s largest retailers, has attracted controversy, with calls for a code of practice to be implemented.

Last month Australia’s Office of the Australian Information Commissioner (OAIC) issued guidance on the use of the technology after a two year investigation found Bunnings Group had breached Australians’ privacy by collecting personal and sensitive information using a FRT system.

Biometric information and templates are considered sensitive information under the Australian Privacy Act and, except for some exceptions, consent is required for the collection of sensitive information.

Bunnings was one of three companies, alongside Kmart, which is also under investigation, and The Good Guys whose use of facial recognition technology was highlighted by consumer advocacy group Choice in 2022.

In September, the OAIC also concluded an investigation into 7-Eleven’s ‘inadvertent reactivation’ of the technology, saying it was satisfied practices and procedures had been implemented to prevent ‘any further recurrence of the conduct’.

A News.com.au poll of nearly 11,000 Australians found 78 percent backed Bunnings use of FRT.

Commissioner Carly Kind says FRT and the surveillance it enables, is ‘one of the most ethically challenging new technologies’.

“We acknowledge the potential for facial recognition technology to help protect against serious issues, such as crime and violent behaviour. However, any possible benefits need to be weighed against the impact on privacy rights, as well as our collective values as a society.”

Kind says while FRT may have been an efficient and cost effective option for Bunnings in its efforts to address unlawful activity, including violence and aggression ‘just because a technology may be helpful or convenient, does not mean its use is justifiable’.

“In this instance, deploying facial recognition technology was the most intrusive option, disproportionately interfering with the privacy of everyone who entered its stores, not just high-risk individuals.”

While Bunnings had argued that it didn’t ‘collect’ biometric information of those entering the store where there was no match, the Privacy Commissioner disagreed, saying that even though the data was deleted within 4.17 milliseconds, it was in that brief time stored on a hard disk, saying it didn’t matter for the purposes of establishing ‘collection’ that the information was held momentarily before being deleted, or that the matching process was conducted automatically and without human intervention.

The OAIC guidance calls for companies to implement a ‘privacy by design’ approach and consider necessity and proportionality, consent and transparency, accuracy, bias and discrimination issues, and governance and ongoing assurance.

The Foodstuffs trial followed a sharp increase in retail crime and acts of violence and aggression in stores.

The latest crime stats from FSNI show there were 40 assaults in the quarter to September 2024, up from 29 a year earlier. Statistics from Woolworths earlier this year shows physical assaults against its employees was up 50 percent year on year.

Foodstuffs North Island’s general counsel, Julian Benefield, says Scarlatti’s conclusion that FRT prevented more than 100 serious events at just 25 stores over a six-month period shows the potential for harm reduction across the wider store network.

“Every case of assault and abuse has an impact on at least one of our team members – they can even be life-changing events. Our stores must be allowed to take the reasonable steps available to them to try to ensure that doesn’t keep happening,” Benefield says.

His views are supported by Retail NZ, whose chief executive Carolyn Young applauded the ‘huge amount of time and effort’ the company has put into the trial ‘to achieve strong outcomes that lead the way for the wider retail sector’.

“Retail NZ members continue to face high rates of violence and crime, putting both their employees and the public at risk, as well as threatening the financial sustainability of retail businesses,” Young says.

“We know that retailers across Aotearoa New Zealand have been watching the trial with great interest and a number are investigating FRT for their own operations in the future.”

She acknowledges, though, that FRT isn’t the solution for all businesses or all crime, but says it’s a tool retailers want to have available to consider alongside other crime prevention tools such as security guards, fog cannons, staff training, body cameras and other tech solutions.

Earlier this year the New Zealand government established a two-year ministerial advisory group – which Young is part of – on retail crime to table proposed legislation for the government.

The results of the Foodstuffs trial are being evaluated by New Zealand’s Privacy Commissioner Michael Webster, with the Office’s own data being reviewed alongside the Scarlatti results.

“At the start of the trial we didn’t know whether FRT would do what the stores wanted, because the evidence wasn’t clear,” Webster says.

“I need to know whether the trial has made a significant difference to the incidence of serious retail crime compared with other less privacy intrusive options.”

Webster is expected to announce his findings before the end of the year.

Foodstuffs will continue using FRT in the 25 trial stores in the interim.

Independent research by OnePicture found 79 percent of the more than 1000 adults surveyed were accepting of FRT being used in retail settings even if the harm reduction was less than one percent, Foodstuffs says.

That flies in the face, however, of an Office of the Privacy Commissioner survey which found 49 percent are concerned about the use of FRT in retail stores.