The Metropolitan Police arrested 173 wanted criminals during a six-month live facial recognition (LFR) pilot in Croydon, running from October 2025 to March 2026. Static cameras mounted on lampposts replaced the usual van-based systems, enabling faster, more agile operations that contributed to a 10.5% drop in crime for the area. The campaign notably cut violence against women and girls by 21%, highlighting the tech’s role in keeping London streets safer.
Record Number Of Arrests
The 24 operations leveraging static facial recognition cameras led to one arrest every 35 minutes. Among those apprehended were suspects wanted for serious offences including rape, kidnap, and sexual assault. Notable arrests included a woman evading capture for over 20 years and a man wanted for voyeurism for more than six months.
Crime Reductions And Hotspot Impact
Crime fell dramatically during the pilot period, with 61% of related offences committed within Croydon itself. The technology effectively focused police resources on local hotspots, preventing repeat offences and ensuring swift justice for serious crimes.
Technology And Public Safety
Static LFR cameras were mounted on existing street furniture, monitored remotely to free up specialist vans. This new setup expanded the Met’s LFR capacity while maintaining operational accuracy. Over 470,000 passersby were scanned with just one false alert, which led to no wrongful arrests.
Ongoing Use And Community Engagement
The Met plans to continue using static facial recognition cameras in Croydon as part of regular deployments. Engagement with local residents and councillors ensures transparency around LFR use, safeguards, and intelligence-led operations designed to balance safety and privacy.
Law Enforcement Statements
“These results show why live facial recognition is such a powerful tool when it’s used carefully, openly and in the right places,” said Lindsey Chiswick, national and Met lead for live facial recognition. “The public can see the difference, with crime down by over ten per cent.”
