‘I Was Scared To Go Into a Shop’: Innocent Shoppers Caught in AI Facial Recognition Nightmare

Facial recognition technology is increasingly being used by UK retailers to identify suspected shoplifters, but what happens when it gets things wrong

facial recognition
A man who visited Home Bargains to buy vitamins says he was wrongly accused of shoplifting after being flagged by facial recognition technology Andres Siimon on Unsplash

A man who went to Home Bargains to buy vitamins says he was wrongly accused of shoplifting by a facial recognition system. His experience is among several cases reported by TechRadar, as Sainsbury's moves to introduce the technology across 200 stores.

A routine shopping trip to buy vitamins left one man afraid to enter shops after he was wrongly accused of shoplifting by a facial recognition system at Home Bargains.

'I was scared to go into a shop,' he told TechRadar.

His wife, identified as Mary, described him as 'distraught' and said the incident left him feeling 'isolated'. The couple were ordered to leave by store management, seemingly without an opportunity to refute the accusation, according to TechRadar.

The incident is one of several accounts detailed in an investigation by TechRadar journalist Samuel Woodhams. Futurism also reported on the cases, drawing on TechRadar's findings.

Shoppers Describe Distressing False Alerts

Another incident involved a woman who was reportedly told to leave a Budgens store after being accused of stealing 'loads of meat' the previous week.

Her husband was left 'shaking from head to toe' after she was told to leave. 'My mind was racing,' the woman said. The couple said they had not visited the store for months.

Other shoppers described being stopped as they entered stores or confronted at tills with full baskets. Some also reported seeing their faces flagged on overhead screens. In one case, a woman was reportedly wrongly intercepted after the system identified people walking into the store behind her, according to TechRadar.

The precise cause of the reported incidents remains unclear. At a meeting in the House of Commons, Labour MP Dr Allison Gardner questioned whether a particular case involved an algorithm glitch, a shop worker making an honest mistake or something else. The accounts therefore raise questions not only about the technology but also about how retailers respond when an alert is generated.

Sainsbury's Facial Recognition Rollout Raises Questions

The concerns have attracted further attention as Sainsbury's is currently in the process of introducing facial recognition technology in 200 of its stores, according to TechRadar.

The publication visited a Sainsbury's store in south London and reported finding no viable way for customers to opt out of the facial recognition system. It also found that security staff had limited information about how the technology worked.

Signs with QR codes were displayed at the store entrance, although people attending the parliamentary meeting said the notices provided insufficient information when disputes arose.

The concerns were discussed at a House of Commons meeting sponsored by civil liberties group Big Brother Watch, where people who said they had been wrongly accused of shoplifting met Dr Gardner, Labour MP for Stoke-on-Trent South.

Neither Sainsbury's nor Facewatch, the company behind the technology, sent representatives to the meeting.

Who Is Responsible When Facial Recognition Gets It Wrong?

Facewatch later said it had given the invitation serious consideration but had concerns about the balance of the event, given the tone of Big Brother Watch's campaign messaging. The company said it had explained its position to the organisers and the parliamentary co-chair, and offered to answer questions after the meeting.

When asked who was liable when the system produced an error, Facewatch said it took responsibility for its technology and processes, while retailers were responsible for how their staff interacted with customers. It added that legal liability depended on the facts of each case and could not be determined through a blanket answer.

The company also said that when a false alert occurred, it initiated a 'safeguarding process with the retailer to establish exactly what happened'.

Facewatch claimed that its technology operated successfully in thousands of stores across the UK every day. It argued that presenting exceptional incidents as representative of its everyday operation was misleading and risked creating unnecessary fear.

TechRadar reported that Booker Group, which owns Budgens, declined to comment on the record. Sainsbury's and Home Bargains did not respond to its requests for comment.

Shoppers Face Difficulties Clearing Their Names

Big Brother Watch helps affected individuals navigate Subject Access Requests (SARs), which allow people to ask organisations what personal data they hold about them under UK data protection law.

However, TechRadar reported that the process could be slow, required affected individuals to provide additional personal information and offered no guarantee that their faces would be removed from watchlists.

Big Brother Watch's head of advocacy, Jack Coulson, described the process as 'the stuff of Kafkaesque nightmares', arguing that people should not have to hand over additional personal data simply to be told the accusation against them.

The UK's Information Commissioner's Office (ICO) acknowledged the severity of the testimonies, calling the experiences 'harrowing'. It said it hoped to share more concrete insights into the commercial use of the technology soon.

Dr Gardner also told TechRadar that she was 'extremely concerned at the poor implementation and lack of governance' surrounding the technology. She said she wanted to know what redress was being given to people wrongly accused, whether it was proportionate to the distress caused and who could be held accountable.

As Sainsbury's moves ahead with its planned rollout, the reported cases leave a central question unanswered: when facial recognition wrongly flags a shopper, who is responsible for putting things right?