On a seemingly unremarkable afternoon in south-east London, Matt Arnold was doing exactly what countless shoppers do every day: picking up supplies for an event, scanning his groceries at a self-checkout, and tapping his Nectar card before paying. The 46-year-old comedy promoter had stopped at the Sainsbury’s superstore in East Dulwich to buy provisions for a standup night at Dulwich Hamlet football club, just next door. It should have been a straightforward errand, the kind of mundane task that barely registers in memory. Instead, it turned into a scene that left him feeling publicly humiliated and profoundly shaken. Within moments of completing his scan, two managers approached him and told him, without preamble or explanation, that he could not be served because of an earlier incident. Arnold was stunned. He had done nothing wrong. He had no idea what they were talking about. But the managers were insistent, telling him he had to leave the premises, and making clear that they intended to escort him out as though he were a shoplifter. For Arnold, the shock was immediate and visceral. He described feeling embarrassed, mortified, humiliated, and powerless. There was no chance to defend himself, no opportunity to ask what he was accused of, no recognition that he might be innocent. The machine had made a judgment, and the people around him were prepared to follow it without question.
Arnold tried to hold on to his dignity as he was effectively being ejected from the store. He refused to be physically walked out, telling the managers that he would leave on his own terms. “They simply told me they would be walking me out,” he later recalled. “At that point I did stand my ground and said I would be walking myself out. I was determined to keep some dignity. On this point, if nothing else, they did relent.” As he made his way out, he glanced up at an overhead CCTV monitor and saw an alert displayed on the screen, a red circle surrounding his face. It was a chilling sight, the visible sign of an invisible system that had labelled him a criminal. The absurdity of the situation was compounded by the fact that he was not behaving like a shoplifter at all. He had a trolley full of supplies for a comedy event, and he was due to be at the football club next door within minutes. Rather than argue further, he asked the staff to keep his shopping safe so that a friend could come and collect it. They seemed confused by the request, perhaps because it did not fit the narrative that had been handed to them. But they agreed, and a colleague named Dave went in a few minutes later to pay for and collect the groceries. As Arnold later reflected, there was no pause for thought from the staff, no suggestion that they understood this was not how a shoplifter would behave. They were simply following the machine’s orders, and the experience left him with a deep sense of powerlessness that stayed with him long after he left the store.
Arnold’s ordeal might be dismissed as a single unfortunate mistake, but it is part of a troubling pattern. Sainsbury’s has been using facial recognition technology from a company called Facewatch, which is designed to identify known shoplifters and other offenders in real time. The stated purpose is to protect staff from abuse and to deter repeat criminals from entering stores. But the system has repeatedly produced false positives, with innocent customers being wrongly identified as criminals and publicly confronted. Last September, another shopper, Warren Rajah, was ordered out of a Sainsbury’s branch in Elephant and Castle after being mistakenly flagged by the same Facewatch software. Similar incidents have been reported in other retail chains, including Home Bargains and B&M stores. Each case follows a familiar pattern: a harmless customer goes about their business, an AI system scans their face, compares it to a watchlist, and produces a match. Staff then act on that match as though it were absolute proof of wrongdoing, subjecting the customer to interrogation and ejection before any meaningful check is made. For Arnold, the fact that this is happening across multiple stores and multiple companies is deeply alarming. He has called on Sainsbury’s, and any other business using Facewatch, to disable the technology immediately, at least until it can be guaranteed to work flawlessly. He has also pointed out that the system is centralised, so if there is a problem in one store, there are potential problems in all of them. Suspending the surveillance in East Dulwich, where his incident occurred, is not enough. The whole system should be paused while an investigation takes place.
Arnold is not simply angry about his own experience, although that would be understandable. He is worried about the broader consequences of allowing machines to make decisions about people’s lives, especially in public spaces where ordinary citizens have no way of knowing they are being watched. He acknowledged that he is not comfortable with surveillance culture, but he can see that it is sometimes necessary. What he cannot accept is the idea that machines should be given the power to make accusations and then expect humans to carry out those accusations without question. “I don’t like surveillance culture, but I can see how at times it is necessary,” he said. “What isn’t necessary is allowing machines to make decisions and then expect humans to scuttle off and do their bidding unquestioningly.” He also expressed concern about what could happen if the technology is turned on people who are less able to defend themselves. “Anyone could be falsely accused,” he warned, “and at some point that will be someone vulnerable, someone with mental health issues like anxiety. It’s inevitable. Also, I would worry about the confidence-destroying effect of it happening to a younger person or someone less willing or able to stand up for themselves as I have done.” His words are a reminder that the cost of algorithmic error falls unevenly. A confident adult with a strong sense of self may be able to absorb the shock and demand an apology. But the same experience could be devastating for someone who is already anxious, isolated, or struggling to cope with daily life.
When Sainsbury’s responded to Arnold’s complaint, the company apologised to him and confirmed that it had paused the use of its AI-assisted Facewatch technology in the East Dulwich store while an investigation took place. A spokesperson said: “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. Customers can be reassured that the Facewatch system has a 99.98% accuracy rate, and every match is reviewed by a trained manager.” Facewatch itself also denied that its technology was at fault. A spokesperson said: “A correct alert was sent to the retailer, but was subsequently subject to human error in the way it was handled in store.” The company described the suspension as a “precautionary measure” to prevent further alerts while staff underwent additional training. Sainsbury’s added that alerts remained visible on staff devices for up to an hour and that it logged all misidentifications. For Arnold, these reassurances ring hollow. The distinction between human error and technology error may be meaningful to lawyers and corporate communications teams, but to the person standing in the middle of a supermarket while two managers order him to leave, it makes little difference. The system as a whole failed him, whether the mistake happened at the point of facial recognition or at the point of human judgment. And the claim of 99.98 percent accuracy is cold comfort to someone who has been publicly branded a criminal. To a statistician, a 0.02 percent error rate might seem negligible. To a human being, it means that thousands of innocent people could be falsely accused every year. It means that somewhere, at some moment, someone is being humiliated in front of strangers because a machine made a mistake.
Arnold’s story is ultimately a cautionary tale about the rush to deploy powerful surveillance tools without fully thinking through their consequences. Technology has enormous potential to make life safer and more efficient, but it also has the power to strip away dignity, disrupt innocent lives, and erode trust in the institutions that are supposed to serve us. The fact that a supermarket chain uses facial recognition to protect its staff from abuse is not unreasonable in itself. But the algorithm is not the final authority. Human beings must remain responsible for the decisions that affect other human beings. When a system flags a customer, there should be careful review, open dialogue, and a willingness to admit uncertainty. There should never be a situation where an innocent person is treated like a criminal simply because a screen displayed a red circle around their face. Arnold has called for the technology to be suspended in every store, not just in East Dulwich, and for the system to be proven flawless before it is used again. That may be a high bar, but it is the right bar. Once surveillance technology is accepted as normal, it is very difficult to push back against its expansion. Every false accusation, every humiliated customer, every quiet surrender in the face of an automated alert is a reminder that the cost of getting this wrong is not measured in percentages but in human suffering. Matt Arnold was fortunate enough to stand his ground, to demand an apology, and to force the company to take notice. But the next person may not be so lucky, and that is precisely why his story matters.