A brand new way of being surveilled could be coming to a store near you—a facial recognition system designed to detect when retail workers have anomalous interactions with customers.
About a month ago, Israel-based Corsight AI began offering its global clients access to a new service aimed at rooting out what the retail industry calls “sweethearting,”—instances of store employees giving people they know discounts or free items.
Traditional facial recognition systems, which have proliferated in the retail industry thanks to companies like Corsight, flag people entering stores who are on designated blacklists of shoplifters. The new sweethearting detection system takes the monitoring a step further by tracking how each customer interacts with different employees over long periods of time.
Shai Toren, Corsight’s CEO, told Gizmodo that the system analyzes how close customers stand to different employees and whether returning customers consistently go to the same employee when they visit a store. Anomalies trigger alerts to store security staff, who decide how to proceed.
“If you go into a shop and you pick up a few groceries, usually you would pick any of the cashiers that is around and you go scan your goods,” he said. “When someone is planning a sweethearting theft, they will always go to the same cashier, which is most of the time a relative of theirs, and this is an anomaly in the behavior compared to the other customers. Our system is able to identify this anomaly and alert on that.”
Advocates for retail workers say the system is based on the misguided assumption that a customer showing loyalty to a particular salesperson is a sign of wrongdoing.
“We have a lot of concerns about this type of technology given that a lot of our members work on commission so the idea is that you are building a book of business based on relationships with customers,” said Chelsea Connor, communications director for the Retail, Wholesale, and Department Store Union (RWDSU). “Whether or not they work on commission, [stores] push sales people to develop those relationships because that’s what brings people back to brick and mortar as opposed to buying online.”
Corsight says that some of its clients are already using the sweetheart detection system but declined to identify them.
Over the past several years, large retail chains have increasingly installed facial recognition and other algorithmic surveillance systems, justifying the increased surveillance by pointing to industry group warnings of “rising” retail crime.
Early coverage of Corsight’s new detection system in industry publications, claims that sweethearting is a growing challenge that contributes to retailers losing $100 billion to theft annually. Those assertions appear to be based on reports from the National Retail Federation, which was forced last year to retract some of the claims it made about the scope of retail theft after an investigation by Retail Dive found that the group’s annual theft analysis was based on a misinterpretation of its own data.
Based on data from its most recent security survey, covering 2022, the NRF says that insider theft, including sweethearting, accounted for 29 percent of inventory losses known as shrink. It said that 3 percent of the retailers included in its data had fully implemented facial recognition systems and another 40 percent were researching or in the process of implementing facial and feature recognition.
The spread of algorithmic surveillance systems in workplaces has prompted federal regulators to warn employers about misusing tools that predict and create dossiers of employee behavior. And last year, the Federal Trade Commission banned the pharmacy chain Rite Aid from using facial recognition after it found that the company’s system had falsely flagged customers, particularly women and people of color, as shoplifters.
Caitlin Seeley George, managing director of the nonprofit Fight for the Future, which has asked retailers to pledge not to use facial recognition, said that in addition to being worried about biases in these systems customers should also be concerned that corporations are inflating fears about theft in order to justify installing surveillance systems that can be used to profile customer behavior for marketing purposes.
“The information retail associations are sharing is cherry picked to make the case for using this technology that they may want to use for a variety of reasons,” she said. “It just opens up the door for mission creep beyond what they’re claiming to focus on.”
Sweethearting detection is just the beginning of Corsight’s work on monitoring not only who is in a store, but how they’re behaving, said Dror Simsolo, the company’s marketing director.
“This is a different flavor of identification,” he said.
English (US) ·