Sainsbury's Paused Facial Recognition at One Store

A muted blue-grey painterly scene of a self-service supermarket checkout aisle under cold fluorescent strip lights, shelves fading into a hazy background.
AI-generated editorial illustration. Not a documentary image.

But left it on in 54 others, with up to 150 more planned

Sainsbury's switched the system off at one shop and left it running in 54 others. Up to 150 more stores are due to get it before Christmas.

How it started

The system is not a smarter checkout

The system at the centre of this is not a smarter checkout. It is live facial recognition, supplied to Sainsbury's by a company called Facewatch, and it does one thing: it compares the face of everyone who walks in against a watchlist of people a shop has flagged, and alerts staff when it thinks it has a match. Sainsbury's brought it in to fight shoplifting, which retailers say has climbed sharply. It is the same face-matching technology that alarmed people when it turned up dormant inside consumer smart glasses — only here it is switched on and pointed at every customer who walks in. A year ago it was a cautious experiment — a trial in two stores, Sydenham in south London and Oldfield Park in Bath, with the promise that any face the software did not flag would be deleted at once. The point of the technology, from the first store, was to turn an ordinary shopper's face into something the shop could check against a list.

What changed since

The trial did not stay small

The trial did not stay small. By this month the cameras were running in 55 branches, and Sainsbury's has told suppliers of hardware it wants up to 150 more installed before Christmas: a rollout, not a pilot. What also grew was the list of people it flagged wrongly. In February a shopper named Warren Rajah was escorted out of the Elephant and Castle branch; Facewatch later confirmed there had been no alert against him at all. Then on 6 August a 46-year-old comedy promoter, Matt Arnold, was refused service at the East Dulwich superstore and told he was linked to an earlier incident. On a nearby CCTV monitor he could see a red circle drawn around his own face. He had done nothing.

Where it stands now

One store switched off, 54 still running

After Arnold's ejection, Sainsbury's suspended Facewatch at the East Dulwich store while it investigates and retrains staff, and only there. The cameras kept running in the other 54 branches, and the plan to expand to more stores has not been withdrawn. Both companies say the technology was not to blame. Sainsbury's said the Facewatch system has a 99.98 per cent accuracy rate and that every match is reviewed by a trained manager, and put the incident down to human error rather than the software. Facewatch said a "correct alert was sent" and that the fault lay in how staff handled it. Neither has explained how a correct alert ends with an innocent man circled in red and turned away, which is the part a shopper would most want answered.

What happens next

Up to 150 more stores before Christmas

The one dated commitment on the table is the retailer's own: up to 150 more stores fitted before Christmas. Against it, Sainsbury's says the East Dulwich investigation and the staff retraining are under way, but has named no date for finishing either, and has not said the rollout waits on the outcome. The regulator is not scheduled to return to this: the Information Commissioner's Office examined Facewatch back in 2023 and, after requiring changes, let it continue. The campaign group Big Brother Watch, which triggered that examination, says the system still should not be running, but nothing new is filed. So the honest state of the next few months is one company expanding a technology on a fixed timetable, and no dated step from anyone with the power to stop it.

Read the full story

Sources and verification

This piece reports what Sainsbury's, Facewatch, the ICO and Big Brother Watch have said, and gives no reader advice on their own legal position if they are wrongly flagged. The accuracy figure of 99.98 per cent is the companies' own claim and is presented as such, not as an independently verified measurement. How a "correct alert" resulted in an innocent shopper's ejection has not been explained by either company in any source read for this article, and no explanation is asserted here.