The Rise of the Surveilled Supermarket: How Facial Recognition is Reshaping Retail Security – and Your Privacy
Over 226 million faces scanned. That’s the number of facial scans conducted during a previous trial of facial recognition technology (FRT) in North Island supermarkets, raising alerts 1742 times. Now, Foodstuffs South Island is expanding this practice, piloting the technology in three Christchurch Pak’nSave and New World stores. But this isn’t just about catching shoplifters; it’s a glimpse into a future where retail spaces are increasingly monitored, raising critical questions about privacy, accuracy, and the normalization of biometric surveillance.
Beyond Shoplifting: The Growing Threat to Retail Staff
The impetus behind this move isn’t simply economic loss. Foodstuffs South Island cites a worrying escalation in “serious threatening behaviours” directed towards staff. The company reports identifying 206 repeat offenders responsible for over a thousand incidents in the past year. Interestingly, a significant proportion of these offenders – six out of ten across the South Island network – are under the age of 18. However, the current trial explicitly excludes minors and “vulnerable people” from its watchlist, a decision the company defends as not diminishing the technology’s effectiveness in targeting adult repeat offenders.
How Does it Work? A Deep Dive into the Technology
The system, integrating with continuously recording cameras, utilizes software from Australian firm Vix Vizion and New Zealand’s Auror. Auror, previously focused on license plate recognition, has recently entered the facial recognition space, acknowledging past limitations in accuracy. The process involves creating a temporary biometric template of every shopper, comparing it to a watchlist of individuals previously exhibiting violent or aggressive behavior. A 92.5% accuracy threshold triggers an alert, prompting two trained staff members to assess the situation – potentially involving observation, contacting police, or asking the individual to leave. Detected matches are stored within the Auror platform for up to seven years, while the actual image is deleted the same day, according to the Privacy Impact Assessment.
The Privacy Balancing Act: Safeguards and Concerns
Foodstuffs South Island emphasizes the safeguards in place, including human oversight and a high accuracy requirement. The North Island trial received a cautious endorsement from the Office of the Privacy Commissioner, acknowledging the “high level of privacy intrusion” but finding safeguards “reduced it to an acceptable level.” However, the very act of scanning every shopper’s face raises fundamental privacy concerns. The data is stored on servers managed by Microsoft Azure Australia and Amazon Web Services (AWS) NZ/Australia, raising questions about data security and potential access. The recent implementation of New Zealand’s biometric processing privacy code, while not directly applicable to this trial, underscores the growing regulatory scrutiny of these technologies.
The Role of Auror and the Expanding Surveillance Ecosystem
Auror’s involvement is particularly noteworthy. While its ‘Subject Recognition’ module doesn’t allow police access, the company’s established relationships with law enforcement through its license plate recognition system highlight the broader trend of increasing data sharing between private companies and authorities. This raises concerns about mission creep and the potential for FRT data to be used for purposes beyond its stated intent. For more information on the implications of data sharing between private companies and law enforcement, see the Electronic Frontier Foundation’s work on surveillance technology.
Looking Ahead: The Future of Retail Surveillance
The Foodstuffs South Island trial is likely just the beginning. As FRT becomes more accurate and affordable, we can expect to see wider adoption across the retail sector – and beyond. This could extend to personalized marketing, loss prevention, and even predictive policing. However, this future hinges on addressing critical questions about transparency, accountability, and the potential for bias. Will shoppers be informed they are being scanned? What recourse will they have if misidentified? And how can we ensure these systems don’t disproportionately target marginalized communities? The debate surrounding facial recognition technology isn’t just about security; it’s about the kind of society we want to build.
The increasing use of retail crime prevention tools, like those offered by Auror, will likely become commonplace. Furthermore, advancements in video analytics, such as those provided by Vix Vizion, will continue to drive the sophistication of these systems. The integration of cloud computing platforms like Microsoft Azure will also be crucial for storing and processing the vast amounts of data generated by FRT. Ultimately, the success of these technologies will depend on building public trust and demonstrating a commitment to responsible innovation.
What are your thoughts on the use of facial recognition in retail spaces? Share your perspective in the comments below!