South Wales Police (SWP) in the UK has continued its use of Live Facial Recognition (LFR), which matches camera footage against a watchlist of suspects in real time, scanning more than 230,000 faces between January and March 2026 alone. While AI-driven mass processing of biometric data aids investigative efficiency, concerns over privacy intrusion and creeping surveillance remain strong.
Facial Recognition · Policing & Surveillance
Your Face, Matched in Seconds — and the Backlash Against It
Police in the UK and Japan now run AI face-matching to identify the missing and chase suspects in minutes — even as the EU, US cities and major tech firms move to rein the technology in.
1–2 min
To return a name, age & medical history from a single field photo (UK OIFR app)
~10M
Records in Japan's national police face database, in operation since March 2020
5 mo.
A woman wrongfully detained in Tennessee (2025) after a false match
The scale of stored faces
A commercial tool's image hoard dwarfs a national police database — roughly 70× larger.
~10M
Japan national police DB
700M+
Clearview AI commercial tool
Four regions, two directions
UK — Deploying
OIFR smartphone app for trained officers; live facial recognition trials at a London station.
Japan — Operating
Nationwide face-matching live since March 2020; links to JR East & Tokyo Metro feeds.
EU — Restricting
2021 proposal to ban real-time facial recognition in public spaces in principle.
US — Pushing back
Cities such as San Francisco ban police use; IBM, Microsoft & Amazon halt sales.
How a field identification happens
Photo & details
gender, est. age
→
Database match
within 1–2 min
→
Identity shared
to paramedics
The promise
Identifies missing or unconscious people fast
Speeds up suspect identification
Non-matching faces flagged "Unknown," not retained
The concern
False positives → wrongful arrests
Broad matching without warrants; opaque watchlists
New AI tracks build, hair & clothing to skirt face bans
Balancing investigative efficiency with human rights remains an unresolved challenge requiring social consensus.
In 2021, JR East halted station face-recognition for released individuals, citing that "social consensus has not been obtained" — even as detection of wanted suspects continued.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…