London rail face-scan trial yields no alert-led arrests

A British Transport Police facial-recognition trial scanned more than half a million faces but produced one incorrect watchlist alert and no arrests caused by an alert, the Guardian reported on September 29.

The six-month pilot covered 18 deployments at busy London railway stations between February and July. Records obtained through a freedom-of-information request put equipment and staffing costs at £320,786 and police time at almost 100 hours.

Why it matters: Rail passengers had their biometric data processed at scale, while the system generated no correct watchlist match during the reported period. That result gives lawmakers and oversight bodies a concrete deployment record for judging whether the intrusion was proportionate.

British Transport Police told the Guardian that officers made other arrests during the deployments for offences including assault, theft and possession of an offensive weapon. Those arrests did not result directly from facial-recognition alerts and therefore do not appear in the system’s performance data.

The force said the pilot was intended to learn how the technology could identify wanted people and individuals who might threaten passengers or staff. It began with a limited watchlist and adjusted locations, operating procedures, equipment and watchlist construction during the trial.

Transport for London supported the extension, saying deployments would target people on police watchlists at selected stations. The public record therefore contains institutional support for continuing the trial despite the first period’s result.

The reported result is narrower than a general finding that facial recognition cannot work. A watchlist system can identify only people included in its reference list who pass a camera under usable conditions. The trial therefore tests the whole deployment design—locations, timing, image capture and watchlist composition—not only the matching algorithm.

The pilot has already been extended

British Transport Police extended the trial for four months and added London Underground stations before the figures became public. The force said the extension generated three confirmed alerts involving people who were complying with sexual-harm prevention orders or other court conditions. Those later alerts fall outside the February-to-July results.

Former UK biometrics and surveillance camera commissioner Fraser Sampson told the Guardian that police must demonstrate proportionality. Sampson, now a non-executive director of retail facial-recognition provider Facewatch, said the outcome depended on the locations, times and watchlists used and called the original trial unproductive.

The policy question is active because more than half of police forces in England and Wales have deployed live facial recognition, according to Liberty Investigates’ analysis cited by the Guardian. London’s Metropolitan Police and mayor have separately announced fixed cameras for the West End.

The source record has limits. The Guardian and Liberty Investigates reviewed the freedom-of-information document, but the underlying document was not publicly linked in the article. The performance figures are therefore independently reported rather than directly inspectable here. British Transport Police’s detailed response is reproduced in the article, not on a separate public results page.

The next evidence will come from the extended pilot. Its watchlist design, confirmed-alert count, resulting police actions and total number of scanned faces will determine whether the initial zero-arrest result reflected poor deployment choices or a persistent mismatch between the technology and the rail setting.

Verification