BTP activates live facial recognition at London Underground stations from October 2024
BTP has extended live facial recognition into the London Underground without releasing performance metrics or new oversight mechanisms. Prior trials showed elevated false positives and retention issues documented by the ICO. The move increases incidental biometric capture in high-density public spaces.
British Transport Police expanded its live facial recognition trial from rail stations into London Underground locations including those served by the Elizabeth line. The deployment uses fixed cameras to capture faces of all passers-by and compare them against a watchlist of persons of interest. No independent accuracy figures were released for the Underground environment.
Prior BTP trials recorded 70 percent false positive alerts according to internal logs referenced in ICO correspondence. NIST FRVT benchmarks show demographic performance gaps exceeding 10 percent for certain cohorts under varying lighting. The Underground rollout adds high-volume, enclosed spaces without new statutory safeguards beyond existing data protection law.
The expansion connects to earlier deployments by the Metropolitan Police and Scottish forces where oversight reports documented retention of non-match images beyond stated deletion periods. Privacy risks concentrate on the absence of mandatory independent audit and the scale of incidental capture in daily commuter flows exceeding one million passengers.
Next steps include planned extension to additional stations and potential integration with transport operator data systems by mid-2025 absent regulatory intervention.
ICO: Formal compliance notice issued within 120 days if non-match image retention exceeds 30 days.
Sources (3)
- [1]Primary Source(https://www.btp.police.uk/news/btp/news/england/btp-expands-live-facial-recognition-lfr-trial-into-london-underground-stations/)
- [2]Supporting Source(https://ico.org.uk/media/about-the-ico/consultation-responses/2023-12/response-to-btp-lfr-consultation.pdf)
- [3]Supporting Source(https://www.nist.gov/programs-projects/face-recognition-vendor-test-frvt)