Court rejects challenge to Met Police live facial recognition use
Preliminary Why: Publisher not yet named · subject to change
Status: no action reported
No source states: loss or records, developer and model, agent tool, could do, deployer.
Severity moderate: Test: data of others touched with a bounded scope stated. More than three million passers-by were scanned in 2025 and 12 false alerts were logged.
The Metropolitan Police uses live facial recognition vans in London. They scan people walking through chosen areas and compare their faces with a database of wanted criminals or missing people. Images that do not match are deleted, and officers double-check any possible match before deciding whether to stop someone. In February 2024 Shaun Thompson was stopped, detained and questioned after the technology matched him to his brother, who was on bail. He and campaigner Silkie Carlo brought a High Court challenge arguing that the use breached privacy and other rights. On 21 April 2026 the court rejected the claims and found the Met's use lawful with adequate safeguards against abuse. Thompson said he intends to appeal. The Met reports more than 2,100 arrests made with the technology since the start of 2024.
What happened
No source tells the story step by step yet.
What was done
No source states an action or fix.
Risks
Read from the sources; not yet reviewed by a person.
Controls
ApplicationEffective Data minimisation
Control failure chain
L1 · ModelProvider not namedModel provider
No source mentions a control at this layer.
Model version not linked to a model card.
L2 · HarnessHarness not namedAgent framework vendor
No source mentions a control at this layer.
L3 · ApplicationDeployer not namedDeployer
Other controls for this risk at this layer, not mentioned in the sources: Sensitive-data checks on outbound data, Human approval for high-impact actions, Output filtering and redaction, People told they are dealing with AI, Risk assessment per deployment.
L4 · Infrastructure / targetOrganisation hit
No source mentions a control at this layer.
Other controls for this risk at this layer, not mentioned in the sources: Bot and scraping defences.
EffectiveNot effectiveNot effective (bypassed)Did not existNot stated in sourcesNot yet checkedAttacker's ownOrder: model · harness · application · target
Controls for this risk: result in this incident
| Control | Part of the system | Result |
|---|---|---|
| Data minimisation | L3 Application | Effective |
| Human approval for high-impact actions | L3 Application | Not stated in sources |
| People told they are dealing with AI | L3 Application | Not stated in sources |
| Output filtering and redaction | L3 Application | Not stated in sources |
| Sensitive-data checks on outbound data | L3 Application | Not stated in sources |
| Risk assessment per deployment | L3 Application | Not stated in sources |
| Bot and scraping defences | L4 Infrastructure / target | Not stated in sources |
Evidence
No source states: a step-by-step account, when it was found, how many records were exposed, the direct loss, a count of the other parties affected, any court case, the cause, what set it off, what was done.
Show the full account: 3 items from the primary sources
What happened, in order
- January The Home Office set out plans to increase the number of vans from 10 to 50 and make them available to all forces across England and Wales.
- December The Home Office said it wanted law enforcement to use facial recognition at significantly greater scale.
Not dated in the sources:
- Shaun Thompson said he intends to appeal the decision.
Show every field, with its source and grade
The AI system
What happened
Impact
Context
Notes on gaps
The sources describe a court ruling on police use of live facial recognition, not an AI agent failure. The sources state no remediations, no loss figures, no record counts and no reversibility. Thompson's planned appeal is unresolved. Claims by the Met and the government are presented as their statements and were not verified.
LR-2026-0520 · version 1 · published 2026-11-01 after automatic checks; not yet checked field by field by a person · JSON