A wanted criminal was apprehended in under a minute using live facial recognition technology. The arrest, which unfolded outside a Barclays branch in Croydon town centre, was the fifth in 45 minutes on a regular Thursday morning as the Metropolitan Police’s controversial live facial recognition (LFR) cameras chalked up another success.
The trap was set at 10am, when clusters of high-definition surveillance cameras mounted on pillars at the junction of Church Street and North End were switched on. Standing nearby was Kevin Brown, a plain-clothes police sergeant. The wanted man unwittingly walked past one of the cameras and instantaneously Brown’s handheld device beeped furiously, displaying a pin-sharp live photo of the suspect alongside a previous custody image, his name, the suspected crime and any warnings about weapons or drugs.
Uniformed officers standing across the road received the same alert and they all dashed to converge on the suspect. He made a vain run for it and then began to fight hard. A pair of officers jumped on his back and several more piled on top as they brought him down, amid shouting and consternation from bystanders. One officer emptied the man’s pockets: a lighter and what looked like a pair of scissors. He was subdued, arrested and taken away in a police van.
Within minutes Brown’s device was buzzing again with matches at the other end of the street. On that weekday morning the AI-enabled system triggered 19 alerts, resulting in nine arrests for crimes including rape, shoplifting and breach of court orders. Another man was stopped, not because he was wanted for a crime, but to check he was adhering to the terms of a court order; a third was held for a while because he had a criminal record and was on probation. “This is a bit over the top isn’t it?” the second man said as three officers crowded him against a wall. He had had no idea LFR was in operation and added: “I’m not coming to Croydon again. It’s mad. I am all registered.”
How live facial recognition operates
The six-month Metropolitan Police pilot of LFR cameras on vans and fixed to lamp-posts saw the technology deployed at the north and south ends of Croydon high street, operating remotely with officers monitoring feeds from a police operations room five miles away in Sydenham. Every face passing the cameras — as many as 5,000 an hour — was scanned and its biometric data streamed live. An AI-powered system supplied by the Japanese tech company NEC checked the data instantly against a database of wanted suspects and people under court orders.
NEC’s face recognition technology has consistently ranked as the world’s most accurate in benchmark tests by the US National Institute of Standards and Technology, with very low error rates. The high-definition cameras capture images that are streamed live; the AI compares biometric templates against the watchlist, and alerts for potential matches are assessed by officers before any action is taken. Biometric templates are immediately deleted if no match is found. Street signs warned that the system was scanning the face of every pedestrian, but the suspect in the Croydon arrest was one of many passersby oblivious to the digital dragnet.
Scotland Yard has trumpeted the effectiveness of the technology at catching people wanted for violence against women and girls. Since the start of 2024, the Met has made more than 2,100 arrests using LFR, including over 100 sex offenders. In one case the cameras detected a registered sex offender, who was required not to be around children, alone with a six-year-old girl. He was arrested and jailed for two years. The Croydon pilot alone resulted in hundreds of arrests — 103 in its first few months — and the Met said a third of those were for violence against women and girls. Between September 2024 and September 2025, LFR deployments across London led to 962 arrests. In the Croydon pilot, 193 of the 249 arrests resulted in charges or cautions. The Met also claimed the pilot led to a 12% reduction in local crime, including retail and violent crime as well as sexual offences.
Accuracy claims and criticism
A widespread public concern is the risk of racial bias, after early models showed concerning results. Critics have cited studies suggesting that LFR can be less accurate for women and people of colour, with some research indicating Black men are more likely to be incorrectly matched. A 2020 report suggested a risk of “anti-Black” racism within LFR development. Big Brother Watch previously stated that 85% of matches recorded by the Met in May 2023 were false, though more recent data shows a significant decrease.
The Metropolitan Police has said independent testing by the National Physical Laboratory (NPL) found that, at the threshold Scotland Yard sets to determine a match, the system was accurate and balanced with regard to ethnicity and gender. An NPL report in March 2023 indicated no statistically significant bias in relation to gender and race at the face-match thresholds the Met was using. The Met said that in 2025 there were just 12 false alerts out of more than 3 million faces captured — a false alert rate of approximately 0.0003% — and that none of these false alerts led to an arrest. However, a separate report in late 2025 indicated that 80% of false alerts in that year were from people with a Black background. The Met has suggested that the location of deployments might explain this overrepresentation.
Elsewhere, Essex Police paused its LFR use after a study found the system was statistically more likely to correctly identify Black people than other ethnicities, although it was still identifying about half of the people on its watchlist. Met Police Commissioner Sir Mark Rowley has said the algorithm is used in a “non-discriminatory way” and “does not perform in a way which exhibits bias,” adding that the force had selected its algorithm “with care.”
The technology’s legality has been tested in court. In April 2026, the High Court upheld the lawfulness of the Metropolitan Police’s 2024 policy on LFR, dismissing challenges brought by community worker Shaun Thompson — who was misidentified by the system — and Silkie Carlo, director of Big Brother Watch. The court found the policy contained “clear, precise and effective safeguards” and that the technology’s use was “in accordance with the law.” The safeguards include officers assessing alerts before action is taken, immediate deletion of biometric data if no match is found, and constraints on who, why and where LFR can be used. There is no specific legislation in the UK governing LFR; its use is regulated by common law and existing statutes such as the Police and Criminal Evidence Act 1984, the Human Rights Act 1998 and the Equality Act 2010. The Met’s current policy is considered to adhere to the College of Policing’s authorised professional practice guidance, unlike the earlier South Wales Police deployment that the Court of Appeal ruled unlawful in 2020 due to insufficient safeguards and excessive officer discretion.
Sir Mark Rowley has described LFR as a “gamechanging” tool that keeps the public safe. The Met views it as cost-effective, increasing efficiency and freeing up officer time for community engagement. The force says 85% of Londoners back the use of LFR to keep them safe, with a later report putting that figure at about 80%.
Public and police perspectives
Critics have called the technology invasive, unregulated and anti-democratic. Shaun Thompson described it as “stop and search on steroids,” arguing it treats innocent people as suspects. Big Brother Watch calls it a “mass surveillance tool” and argues it has “no place” at cultural events. Some campaigners point out that LFR has not been voted on in Parliament and operates in a legal grey area, lacking explicit legal authorisation. Similar systems are used in China, the UAE, India and Israel, while the EU’s Artificial Intelligence Act heavily restricts LFR use in public spaces for law enforcement, requiring strict necessity and judicial authorisation.
Public opinion in Croydon town centre ran the gamut from strong opposition to strong support. “It’s not good to have your face scanned,” said Maleek Ife, 36, a delivery driver. “It’s a violation of privacy. To scan everyone? I’m not happy with that. It’s not what I believe the UK is about. You should respect people’s privacy. You should be able to walk around freely. I don’t think we should become a surveillance state. We are going to become like the countries we criticise.”
“It’s good,” said Sam Mensah, 53, a supermarket worker. “If someone is doing something bad you can catch that person immediately. I am not worried about being scanned. I have no issue to hide.”
Owen Brown, 63, a carer, saw the cameras as just another part of a wider slide into digital tracking. “The way life is moving now they track you through your phone anyway,” he said. “There’s nowhere you can go without being scanned or looked at. It’s invasive, but what can you do about it? It’s part of life now.”
