Sainsbury's has temporarily stopped using its AI-powered facial recognition system at its East Dulwich store following an incident where a customer was wrongly accused of shoplifting and asked to leave. The supermarket chain said the error was due to human misjudgment by staff, not a fault with the Facewatch technology itself.
The customer, Matt Arnold, 46, was at a self-service checkout on August 6 when store management approached him. They informed him he could not be served due to an alleged incident earlier in the week. Arnold, who was purchasing items for a comedy event, was asked to leave the store. As he exited, he noticed an overhead CCTV monitor displaying a red circle around his face. Sainsbury's head office contacted Arnold the following day to offer an apology.
Sainsbury's stated that its Facewatch system has a 99.98% accuracy rate and that every alert is reviewed by a trained manager before any action is taken. The company maintains that the technology correctly identified Arnold but that store staff mishandled the alert. Facewatch also defended its technology, asserting that a correct alert was sent but handled improperly by the retailer. The suspension at the East Dulwich store is described as a precautionary measure while staff receive additional training.
This is not the first instance of a customer being wrongly identified by Sainsbury's facial recognition system. In February, Warren Rajah was escorted from the grocer's Elephant and Castle branch after being mistaken for an individual flagged by Facewatch. At that time, Facewatch confirmed Rajah had no associated alerts or incidents, and Sainsbury's attributed the error to human error.
Sainsbury's has been expanding its use of facial recognition technology, with plans announced in July to implement Facewatch cameras in up to 150 additional stores before Christmas, adding to the 55 branches where it was already in operation. The company stated that trials of the system had resulted in a 46% reduction in theft, harm, aggression, and antisocial behaviour, with 92% of identified offenders not returning to its stores. Facewatch's system compares shoppers' faces against a watchlist of individuals previously linked to theft, violence, or aggression.
Privacy and civil liberties groups have voiced concerns regarding the increasing use of biometric surveillance in retail and the potential for false identifications. Silkie Carlo of Big Brother Watch noted that even a small error rate, when applied to millions of weekly shoppers, can lead to a significant number of misidentifications.
The company has previously stated that the Facewatch system is designed solely to identify individuals who have previously shown violent, aggressive, or criminal behaviour in participating stores and is not used to monitor everyday shoppers. Images are deleted instantly if the system does not recognise someone.