Sainsbury’s branch halts AI use as shopper ejected

Sainsbury’s Pauses Facial Recognition at London Store After Wrongful Ejection

Ninoda.com – A south-east London Sainsbury’s superstore has temporarily halted its live facial recognition system following an investigation into a customer who was wrongly identified as a shoplifter and ordered out of the premises. The technology, supplied by Facewatch, had been trialled at select stores since the previous year.

The Incident

On 6 August, Matt Arnold, a 46-year-old comedy promoter, was at the East Dulwich branch purchasing supplies for a stand-up event hosted at the adjacent Dulwich Hamlet Football Club. After scanning his items and his Nectar loyalty card at a self-service till, he waited for a staff member to authorise an alcohol purchase and waved an attendant over.

Rather than receiving service, Arnold was intercepted by two members of management who informed him he could not be served owing to an incident “earlier in the week.” He was then escorted from the store.

“They came over and said I had to leave. The staff member said I’d been identified by the AI, and the cameras had flagged me.”

Arnold noted that a genuine shoplifter would not typically carry a large basket of goods through the scanner or present a loyalty card. From the entrance, he later spotted an overhead CCTV monitor displaying a red circle around his face.

“But what upset me was thinking this is what the future could be – people just listen to what the machine tells them to do without thinking about the consequences.”

Company Response

The following day, Sainsbury’s head office reached out to Arnold with an apology, attributing the mix-up to “human error” rather than a software malfunction. A spokesperson confirmed the system would remain paused at the Dulwich location during the investigation while defending the broader rollout, pointing to rising violence against retail workers as justification.

“We have contacted Mr Arnold to apologise for his experience at our Dulwich superstore. The incident was caused by human error, not the facial recognition technology.”

The spokesperson added that the Facewatch platform carries a 99.98% accuracy rate and that every match is reviewed by a trained manager before action is taken.

A Pattern of Errors

This was not the first time the Facewatch system deployed at Sainsbury’s stores has produced a false identification. In September, Warren Rajah was ordered out of an Elephant and Castle branch after being wrongly flagged by the same software; the security firm subsequently confirmed he was not even registered on its database. That episode, too, was attributed to “human error” by both Sainsbury’s and Facewatch.

Facewatch has also been linked to comparable incidents at other retail partners, including a Home Bargains store in 2024 and a B&M in Cardiff the previous year.

A Facewatch spokesperson stated their technology was “not at fault” in the Arnold case, explaining that a correct alert had been transmitted to the retailer but was then mishandled in-store. The company described the store-level suspension as a “precautionary measure” to prevent further alerts while staff receive additional training.

Arnold’s Perspective

Arnold, who described the ordeal as a “terrifying glimpse of the future,” expressed sympathy for the floor staff, believing they were ill-equipped to operate the system amid the pace of a busy store.

“Even though what happened to me was horrendous, I feel sorry for the staff. Mistakes are going to happen. They’re not being equipped to implement the AI properly.”

He also challenged the framing of the error rate. Even accepting Sainsbury’s claimed 0.2% inaccuracy, he argued that figure translates into a substantial number of wrongly targeted shoppers given that millions pass through Sainsbury’s stores every week.

“There are more people than me who have been caught out by this.”