Facial recognition trial at London stations yields no matches in six months

A trial of live facial recognition cameras at London railway stations has failed to produce a single confirmed match in six months, British Transport Police (BTP) has confirmed. The technology, deployed at Victoria Underground station since February and set to rotate between other stations until November, scans faces in real time against a watchlist of individuals wanted for serious offences, including robbery, knife crime and sexual assaults.

BTP said no alerts from the trial have resulted in a confirmed match. The force has set a high threshold for accuracy to reduce the risk of misidentification. While no matches have been made, officers have made 13 arrests during deployments for unrelated offences, such as breach of Criminal Behaviour Orders and theft. Four of those arrested were wanted by other police forces.

Public concerns over privacy and accuracy

The trial has raised concerns among Londoners about privacy, racial bias and the potential for errors. An Opinium poll, commissioned by facial recognition firm Face Int, found that 72% of Greater London respondents want public input on how the technology is used. Two-thirds of those polled worry that errors could lead to innocent people being wrongly accused.

Tony Kounnis, CEO of Face Int, said the public’s stance on facial recognition is nuanced. "People have specific and understandable concerns, particularly the accuracy of the technology, their own privacy, and the security of biometric data," he said, as reported by MyLondon. "Public trust cannot be overlooked; nor is the efficacy of the technology enough."

Green Party London Assembly Member Benali Hamdache told the Local Democracy Reporting Service (LDRS) that stronger protections are needed to prevent overreach. "With instances of misidentification and a lack of transparency surrounding how data is used, it’s clear that proper oversight is urgently needed," he said.

Jasleen Chaggar, Senior Legal and Policy Officer at Big Brother Watch, criticised the lack of public consultation. "Londoners understand the risks of error-prone AI," she said, as reported by MyLondon. "They can put themselves in the shoes of someone who is falsely flagged by a machine, then accused and questioned by the police and forced to prove their innocence."

What happens to the data

BTP has said that images of faces not matching the watchlist are deleted immediately and not retained. The force added that the trial is part of wider policing operations, including patrols and engagement activity, and that crime has reduced during periods when the technology has been in use.

A BTP spokesperson said the trial aims to assess how the technology performs in a railway environment. "LFR is being deployed as part of wider policing operations at stations, alongside existing patrols and engagement activity, to keep passengers and staff safe," they said. "As these arrests were not as a direct result of a facial match using the LFR technology, they are not included within LFR performance data."

The Mayor of London’s Office for Policing and Crime (MOPAC) said it will continue to scrutinise the use of live facial recognition. A spokesperson for the Mayor told the LDRS: "The Mayor is committed to ensuring that police are open and transparent around where, when, why and how Live Facial Recognition technology is used in London."

What happens next

The trial will continue until November, with the technology rotating between different London stations. BTP has not confirmed which stations will be next. The Metropolitan Police is also planning to use static facial recognition cameras in London’s West End and Soho, areas with high crime rates. The Mayor of London, Sadiq Khan, appeared to confirm its use on Oxford Street during a City Hall session last month.

Campaign groups, including Big Brother Watch, have called for the government to adopt safeguarding proposals and allow a public debate on the use of facial recognition technology. It remains unclear whether the trial will be extended or expanded beyond November.