Analysis: London police used facial recognition 117 times in 2024 to August, up from 32 times total from 2020 to 2023; ~770,966 people were scanned since 2020
Yasemin Craggs Mersinoglu / Financial Times :
Context & Ripple Effects
London had already been positioned as a test bed for visual-surveillance technology amid a large CCTV footprint. Police then set out a model of time-limited live facial-recognition deployments using tailored suspect lists in their 2020 rollout plan.
The reported 2024 increase shows that model moving from occasional use toward a far more frequent operational tool, making the scale of biometric scanning a central public-safety AI governance question.
First-order effects
- London police are using live facial recognition far more often than in the preceding four years combined, expanding the number of encounters in which people in monitored areas are scanned.
- The approximately 770,966 scans recorded since 2020 make the system's operational footprint more visible to residents, civil-liberties groups, and policymakers evaluating its use.
Second-order effects
- More frequent deployments increase pressure for clear rules on watchlist construction, oversight, data handling, and how police demonstrate that use is proportionate.
- The acceleration gives practical weight to the UK’s earlier plan to expand police biometric systems, potentially making London’s deployment experience a reference point for other forces.
Third-order effects
- If high-frequency deployment persists, live facial recognition could become a routine layer of urban policing rather than an exceptional tactic, shifting governance debates from authorization to continuing constraints and accountability.
- The case illustrates a broader contest in public-sector AI: operational adoption can advance quickly once camera infrastructure and police workflows are in place, while safeguards must keep pace with scale.
The trend: Police forces are moving biometric AI from limited trials to repeatable operational deployments, intensifying demands for durable public-safety AI governance.