British Transport Police's six-month live facial recognition pilot on London's rail network scanned more than half a million faces and generated exactly one alert, which turned out to be a false positive. According to figures obtained by civil liberties group Liberty via Freedom of Information requests and reported by The Guardian, the trial cost over £320,000 and consumed nearly 100 hours of officer time, with no arrests directly resulting from a live facial recognition alert.
The system, built on NEC's NeoFace M40, compared passers-by against a police watchlist and flagged potential matches for human review. BTP says it deletes biometric data of non-matches and that deployments are intelligence-led, targeting crime hotspots. Privacy groups remain unconvinced; Big Brother Watch's Jasleen Chaggar called the results laughable but troubling, while Privacy International's Sarah Simms said the trial shows how invasive and disproportionate the technology is, especially without a specific legal framework.
Despite the results, BTP has extended the trial until November and expanded it from Network Rail stations to the London Underground. The pilot's poor performance echoes earlier concerns about police facial recognition, including temporary suspensions over racial bias testing. For now, the only person the cameras picked out was the wrong one.