logo
Kurashizu Blogwhere ideas flow
  • Home
  • News
  • Blog
  • About
  • Home
  • News
  • Blog
  • About

© 2026 Kurashizu. All rights reserved.

Admin·New Post·Service Status
Powered by Cloudflare·30.10.26
← Back to news

500k facial scans at UK stations yield no arrests, 1 false positive

Original article
theguardian.com|398 points|239 comments|by ilamont|Sep 29, 2026

Failure of Surveillance: London's LFR Trial Results

A six-month pilot program utilizing LFR (Live Facial Recognition) technology across several of London's most crowded transit hubs has yielded negligible results despite significant expenditure.

People on the platform at King's Cross railway station in London

The Cost of Inefficiency

Data obtained via a freedom of information request by Liberty Investigates reveals a stark disparity between the resources invested and the actual outcomes. Between February and July, the British Transport Police (BTP) scanned over half a million faces.

Trial Metrics Summary

MetricValue
Total Faces Scanned500,000+
Total Cost£320,786
Police Man-Hours≈100\approx 100≈100 hours
Successful ArrestsSome 0
False Positives1

The financial inefficiency can be expressed as: Cost per Alert=£320,7861 alert=£320,786\text{Cost per Alert} = \frac{£320,786}{1 \text{ alert}} = £320,786Cost per Alert=1 alert£320,786​=£320,786

How the Technology Operates

The LFR system does not record everyone; it compares live feeds against a specific "watchlist."

Technical Implementation

The system processes what is known as special category personal data. A hypothetical watchlist entry might look like this:

{
  "subject_id": "BTP-99283",
  "status": "Wanted",
  "offence": "Breach of Court Order",
  "priority": "High"
}

Expansion and Justification

Despite the initial lack of success, the BTP has extended the trial by four months and expanded its reach to include London Underground stations.

  • Recent Developments: Since the extension, BTP reports three positive alerts. However, these individuals were found to be complying with their court-imposed conditions (such as sexual harm prevention orders).
  • TfL's Stance: Transport for London argues that LFR is essential for:
    • Identifying high-risk individuals on police watchlists.
    • Combating violence and sexual harassment against women and girls.

Critical Perspectives

The Expert View

Fraser Sampson, a former biometrics commissioner and current director at Facewatch, emphasizes that the utility of the tool depends entirely on the context.

"We know the technology works, but using it in supermarkets to deter shoplifters... is very different from deploying it to catch people on a public transport network."

Sampson argues that for the police, success is binary: people are either caught or they are not. In this instance, the trial was not fruitful. He stresses that the use of such technology must be appropriate, proportionate, and necessary.

The Human Rights Concern

The rollout has faced significant political pushback:

  • Parliamentary Joint Committee on Human Rights: Identified LFR as a "particularly clear example of risk" regarding AI and human rights.
  • Bell Ribeiro-Addy (Labour MP):

    "Live facial recognition technologies are a serious invasion of our privacy. The government should be suspending their rollout until there are stronger safeguards in place..."

The Broader Surveillance Landscape

The London trial is part of a wider trend. According to Liberty Investigates, more than 50% of police forces in England and Wales have now utilized LFR.

  • West End: Fixed cameras are being installed.
  • Oxford Street: Mayor Sadiq Khan confirmed the technology will be used on the newly pedestrianized stretch.

Surveillance Context

BTP's Defense

A spokesperson for the BTP maintains that the pilot is a learning process. They claim that while LFR didn't trigger the arrests, officers present during the deployments still made arrests for:

  • Assault
  • Theft
  • Possession of offensive weapons
  • Public order offences
  • Breach of criminal behaviour orders

The BTP is currently working on the following refinements:

  • Optimizing deployment locations.
  • Updating operating procedures.
  • Improving equipment.
  • Refining the construction of watchlists.

They conclude that the growth of their watchlists has been "deliberate, controlled and aligned to BTP policy" to ensure data quality and governance.