\u201cWhile six out of the top 10 offenders are minors across the FSSI store network, no minors or vulnerable people will be included in the watchlist during the project,\u201d said the 26-page report, recently made public.<\/p>\n
RNZ asked what percentage of serious threatening behaviour incidents were being committed by minors at its supermarkets, but the company did not say.<\/p>\n
Foodstuffs SI had justified the move into facial recognition in part by pointing to an \u201cescalating level of serious threatening behaviours by repeat offenders in FSSI supermarkets [that] poses a significant risk to the health and safety of staff and customers in-store\u201d.<\/p>\n
The company said the preponderance of teenagers among offenders did not undermine its expectation that adult-only facial recognition was a good way of combating the issue.<\/p>\n
\u201cExcluding minors does not reduce the usefulness of the trial,\u201d said a spokesperson.<\/p>\n
\u201cIts purpose is to assess whether facial recognition technology can help our teams identify adult repeat offenders who have previously been involved in serious incidents, so they can step in early to prevent further harm. The trial is also helping us understand the operational impacts and processes involved with using facial recognition in our stores.\u201d<\/p>\n
It had identified 206 repeat offenders at 38 stores in 32 suburbs in the past year involved in over a thousand events, the privacy assessment said.<\/p>\n
\u201cWe\u2019re seeing increasing levels of serious, threatening and violent behaviour in our stores from repeat offenders across all different age groups.\u201d<\/p>\n
Camera on continuous record<\/p>\n
The three-month trial began in October at its Pak\u2019nSave stores in Redwood and Sydenham, and its New World St Martins.<\/p>\n
The software integrates with a camera that is continuously recording images.<\/p>\n
In an earlier trial in some North Island Foodstuffs stores, the system scanned 226 million faces to raise 1742 alerts, with 1200 of those confirmed on watchlists.<\/p>\n
A watchlist only had on it people who had previously been violent, threatening or aggressive in their stores \u2013 and were then categorised in one of two ways: \u201cExtreme Threat\u201d and \u201cHigh Risk\u201d. A match with someone\u2019s template had to be 92.5% accurate for it to be flagged, at which point two trained store members would check it and decide what to do next, said Foodstuffs South Island.<\/p>\n
\u201cThis may include observing the person, contacting police, or intervening to ask them to leave the FR Store \u2013 if it is safe to do so,\u201d said the chain\u2019s website.<\/p>\n
Facial recognition was a proportionate response because the North Island supermarket trial had been effective at reducing serious threatening behaviour, the privacy assessment said.<\/p>\n
The North Island trial was assessed by the Office of the Privacy Commissioner, which concluded, \u201cWhile the level of privacy intrusion was high because every visitor\u2019s face is collected, the privacy safeguards in the trial reduced it to an acceptable level.\u201d<\/p>\n
The tech must only be used with robust processes including human checks, it said.<\/p>\n
\u201cFRT [facial recognition technology] systems have potential safety benefits, but they also raise significant privacy concerns.\u201d<\/p>\n
A template of every shopper<\/p>\n
The privacy assessment for the current South Island trial showed Foodstuffs was using IT systems from an Australian firm, Vix Vizion, which scored highly for accurate facial recognition in 2022, and Auckland company Auror.<\/p>\n
Auror is a big player in vehicle licence plate recognition but earlier said it had avoided facial recognition until a few months ago because the technology had not been accurate enough at the time.<\/p>\n
Its product, Subject Recognition, had a module and person of interest lists that could not be accessed by police, whereas police access its licence plate system hundreds of times a day.<\/p>\n
Like Subject Recognition, which Auror has begun offering stores, Foodstuffs\u2019 trial runs on a system that created a temporary biometric template of every shopper at the three test stores, then dumped it if there was no match to a watchlist, the privacy assessment showed.<\/p>\n
A detection was stored within the Auror platform for seven years, though the image itself would be deleted the same day, said the assessment.<\/p>\n
\u201cAll personal information collected and stored within the FRT system, including the watchlist, is stored in Microsoft Azure Australia (by Auror Limited within the Auror Platform) and Amazon Web Services (AWS) NZ\/Australia\u201d by Foodstuffs.<\/p>\n
The company said before starting the trial it engaged with the Office of the Privacy Commissioner (OPC). The OPC would be getting \u201chigh-level\u201d monthly updates.<\/p>\n
Facial recognition relies on biometrics, like a person\u2019s face or fingerprints.<\/p>\n
The office\u2019s biometric processing privacy code came into force on November 3 so does not apply to the supermarket trial.<\/p>\n
– RNZ<\/p>\n","protected":false},"excerpt":{"rendered":"\u201cWhile six out of the top 10 offenders are minors across the FSSI store network, no minors or…\n","protected":false},"author":2,"featured_media":200466,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[5],"tags":[192,138,103,2514,111,139,69,31341,21037,31340,1560,418,1240,7227,47066,1555,61,8598],"class_list":["post-200465","post","type-post","status-publish","format-standard","has-post-thumbnail","category-business","tag-are","tag-business","tag-facial","tag-island","tag-new-zealand","tag-newzealand","tag-nz","tag-pennington","tag-phil","tag-recognition","tag-report","tag-shows","tag-south","tag-supermarkets","tag-teenagers","tag-threat","tag-to","tag-worst"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts\/200465","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/comments?post=200465"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts\/200465\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/media\/200466"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/media?parent=200465"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/categories?post=200465"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/tags?post=200465"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}