British Transport Police Trial Scanned 500,000 Faces, Produced No Direct Arrests in First Phase

British Transport Police scanned more than 500,000 people with live facial recognition during the first phase of a trial at London rail stations, but the system produced no direct arrests, according to figures obtained through a freedom of information request.

Officers conducted 18 deployments between February and July. The force recorded one false alert in that period and spent £320,786 on the trial, which also used almost 100 hours of police officers’ time. Arrests made by officers while cameras were deployed should not be confused with arrests resulting from a facial recognition match: none of the latter occurred during the initial phase.

The technology compares faces captured by cameras against a police watchlist and alerts officers to a possible match. Officers must then decide whether to stop the person and verify their identity. The cameras do not themselves establish that someone has committed an offense.

British Transport Police expanded its trial to London Underground stations as it tested the use of the technology across the transport network. BTP said it began with a limited watchlist and refined deployment locations, equipment and operating procedures as the pilot progressed. Its stated aim was to identify wanted offenders and people who may pose a risk to passengers or staff.

The force has reported three confirmed alerts in a later, extended phase of the trial. Those individuals complied with orders, according to the force. That subsequent result does not change the first phase’s tally of zero direct arrests, and the two periods should not be combined when assessing the initial deployments.

Police forces elsewhere in England have reported different outcomes from their own deployments. Thames Valley Police recorded three outcomes from alerts in Slough, illustrating how results depend on the watchlist, location and operational purpose rather than the number of passersby scanned alone.

Former UK biometrics commissioner Fraser Sampson said results depend on where and when cameras are deployed, how watchlists are assembled and whether listed people are likely to pass through a location. BTP said officers made other arrests during the initial deployments, but none stemmed directly from a facial recognition alert.

Sources: The Guardian

–

By the ID Tech Editorial Team