Foodstuffs’ Use of Facial Recognition Offers Ammo to Both Sides of Debate

Recent developments from Foodstuffs, a major supermarket cooperative in Australia and New Zealand, have highlighted the use of facial recognition technology in 25 of its North Island stores—and ignited debate. This trial is part of a broader trend of cautious yet growing acceptance of facial recognition technology in retail.

According to an independent evaluation by analytics firm Scarlatti, the six-month trial, conducted from February 2024, resulted in a 16 percent reduction in “serious harmful behavior” such as assaults, abuse, and disorderly conduct within those stores. During the trial, there were 1,742 facial recognition alerts for repeat offenders or their accomplices, with staff choosing to approach those flagged individuals about half of the time.

Scarlatti’s report indicated that approximately half of the 16 percent harm reduction could be linked to staff approaches, while the other half was attributed to a deterrent effect, presumably caused by the knowledge that facial recognition was in use. Foodstuffs North Island general counsel Julian Benefield noted that more than 100 serious harm incidents had been prevented at just 25 stores over half a year. “We have a moral and legal duty to do all we can to keep our teams and customers safe,” he said.

However, the trial also recorded instances of misidentification. According to the evaluation, nine shoppers were incorrectly flagged, and in two of those cases, customers were asked to leave. Foodstuffs acknowledged these incidents and stated they were due to human error in interpreting the data, not a fault in the facial recognition software itself. The company said it has apologized to those affected and implemented additional training and verification steps to reduce such errors in the future.

A separate development involved a Rotorua woman who was wrongly identified as an offender. She reported feeling embarrassed and “powerless” when approached and asked to leave in front of her child. Now pursuing a case at the Human Rights Review Tribunal, she questions the technology’s fairness and potential biases. This incident is particularly concerning given that Privacy Commissioner Michael Webster has expressed concerns about potential biases affecting minority groups, such as Māori, Pasifika, Indian, and Asian customers. While Foodstuffs maintains that the benefits outweigh the drawbacks, the woman’s experience illustrates the immediate personal consequences when the system fails.

A public survey by OnePicture showed that around 79 percent of respondents would accept the use of facial recognition technology even if it only slightly reduced harmful incidents, suggesting a degree of public tolerance for this security measure. At the same time, the woman’s case and other misidentifications highlight concerns about individual rights and potential discrimination. The differing perspectives underscore the key questions surrounding facial recognition: balancing potential safety benefits against privacy risks and the possibility of error.

Foodstuffs will continue using the technology in the 25 trial stores for now, pending the outcome of the Privacy Commissioner’s inquiry. That inquiry is expected to further evaluate the legal and ethical implications. Meanwhile, the incidents of both harm prevention and misidentification within the trial period have offered a clearer, fact-based picture of what is at stake.

Sources: RNZ, Rotorua Daily Post

–

December 6, 2024 – by Ji-seo Kim