The Facial Recognition System

The facial recognition system Facewatch, already used by over 100 British businesses to monitor shoplifters, is preparing for a leap that has not yet been made in the UK: to notify the police directly, rather than just store staff, when someone deemed to be among a list of “dangerous offenders” appears in front of a camera. This feature will debut in autumn and is presented as a first for the country.

In an illustration of the mechanism to the Guardian, Facewatch CEO Nick Fisher describes it as a “unique technical development”: when the network flags the “worst offenders,” the alert would reach law enforcement in an average of four seconds. The technology is already adopted by brands such as Sainsbury's, B&M, and Spar, and its reach is set to expand rapidly: Sainsbury's alone has announced plans to increase from 55 to over 200 stores covered by the end of the year.

In the first six months of 2026, Facewatch would have alerted retailers almost 300,000 times that a “known offender” had entered a store, a stream of reports that the company claims allows staff to intervene “before theft, assault, or violence could occur or escalate.”

Civil Rights Groups' Objections

Doesn’t this remind you of a movie you’ve seen before? Civil liberties organizations speak of a “dangerous escalation” towards surveillance and criminalization in retail. Charlie Whelton, head of policies and campaigns at Liberty, considers the development “unproven and opaque” and denounces how facial recognition has been allowed to “proliferate without any regulation.”

The legal point raised is clear: “It's not against the law to enter a store even if crimes have been committed in the past. The idea of calling the police on someone who hasn’t committed a crime but is feared to might commit one subverts our way of doing things.” Whelton adds the issue of reliability: “It’s not infallible. These systems make mistakes, and it’s very difficult to contest them when the mistake happens to you.” Several people have been forced to leave a store after being wrongly identified as shoplifters, an experience some have described as “Orwellian,” expressing that they felt “guilty until proven innocent.”

Even those studying the social impact of artificial intelligence question the proportionality of the measure. Nuala Polo, head of British public policies at the Ada Lovelace Institute, notes that “there are other means, much less intrusive, to identify shoplifters, without the need to scan millions of faces every day, practically without consent.”

The group Big Brother Watch is even harsher, criticizing the police for “inserting themselves into this cowboy operation” and warning that citizens would end up facing a “secret blacklist, compiled by companies and private security guards with no form of accountability.”

The crux of the matter remains the lack of regulatory framework: the technology spreads in retail while there is no discipline that specifies who ends up on surveillance lists, with what guarantees, and with what possibility to contest an error. With the direct involvement of law enforcement starting in autumn, the weight of that gap grows.