London Police Facial Recognition Test Results in 173 Arrests
London police tested an AI facial recognition system over six months in Croydon, identifying over 2,000 people with criminal records and leading to 173 arrests with one false identification among 470,000 scans. The department announced plans to expand permanent cameras to central London by late 2026, prompting civil rights groups to warn of privacy risks and potential for racial discrimination.
What happened
The Metropolitan Police tested a real-time AI-based facial recognition system in Croydon, south London, over six months. The system identified over 2,000 people with criminal records, resulting in 173 arrests including a woman wanted for over 20 years, with only one false identification among approximately 470,000 faces scanned.
Crime in the monitored area fell by 10.5% and violence against women and girls declined significantly during the trial. The Met Police announced plans to install permanent cameras in central London locations including West End and Soho by Christmas 2026, with expansion to six additional areas in 2027.
Reported impact
Civil rights organizations including Big Brother Watch raised concerns about mass surveillance of hundreds of thousands of citizens without judicial justification. The group warned of privacy risks, risks of misidentification, and racial discrimination linked to the algorithm.
Index analysis
The system achieved a 99.99% accuracy rate based on one error among 470,000 scans. The Met Police stated the technology assists officers rather than replacing them and cited crime reduction as evidence of effectiveness.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.