How Live Facial Recognition Technology is Being Deployed by UK Police

The Labour government considers facial recognition technology to be the most significant breakthrough for law enforcement since the introduction of DNA matching. As part of this initiative, the government has announced the deployment of 40 new police vans equipped with live facial recognition (LFR) cameras in town centers across England and Wales.

Proponents argue that the technology streamlines police operations and aids in the apprehension of criminals. Conversely, opponents express concerns regarding potential violations of civil liberties and the risk of bias against minority ethnic groups.

Retrospective facial recognition technology, which compares images from CCTV, mobile phones, and other sources against police databases, is already available to all forces in England and Wales. This process involves measuring facial landmarks to confirm identities during investigations.

Live facial recognition (LFR) represents a more active form of surveillance. It allows police to scan faces in real time and compare them against watchlists of wanted individuals. These systems are currently deployed via marked police vans or, in some trial cases, fixed to infrastructure like lamp-posts.

When a match is identified, the information is sent to officers nearby, who then assess the validity of the match before taking action. These cameras are typically operated in short intervals in high-traffic areas or during major events.

Data that does not result in a match is permanently deleted. However, the scale of data collection is significant; for example, nearly one million faces were scanned at the British Grand Prix at Silverstone since 2023, though the operation resulted in zero alerts.

Currently, 13 police forces in England and Wales utilize LFR. The Metropolitan Police in London is the largest user, having scanned over 6.6 million faces since April 2023. Other forces, including South Wales, Essex, and Surrey, have also reported varying levels of usage and arrest rates.

Regarding racial bias, historical studies indicated higher error rates for minority ethnic backgrounds, though recent data suggests improvements. A study of Essex police deployments found the system was statistically more likely to correctly identify black participants than those from other groups, though discrepancies remain in other contexts. Additionally, concerns have been raised about the disproportionate deployment of the technology in areas with higher black populations.

Oversight is currently provided by a variety of bodies, including the Information Commissioner and the Equality and Human Rights Commission. The government is currently consulting on a new, more unified legal framework to govern these practices.

Future developments may include operator-initiated facial recognition, where officers use mobile devices to scan individuals in specific circumstances, such as when a subject is unconscious or refuses to provide details. Furthermore, there is interest in exploring technology that analyzes body movements and facial expressions to infer behavior or emotional states.