Essex Police says AI facial matching drove 57 arrests and zero false positives

Essex Police says its redeployed live facial matching program screened more than 353,000 faces across 22 deployments, producing 173 alerts, 143 interventions and 57 arrests with no false positives. The force says the results show how independently evaluated AI can speed frontline decisions while reducing manual review and unnecessary public encounters.

Why it matters: - Essex Police’s live facial matching program is being presented as a real-world test of whether AI can improve frontline policing without adding risk. - The force says the system produced measurable operational gains, including arrests and interventions, while recording zero false positives. - The results matter because UK police forces are under pressure to adopt AI responsibly, transparently and with public oversight.

What happened: - Essex Police redeployed Corsight AI’s facial intelligence technology across its live facial recognition program after the UK’s National Physical Laboratory evaluation. - Since redeployment, the force has run 22 operational deployments. - Officers screened more than 353,000 faces during public policing operations. - The program generated 173 positive alerts, 143 positive interventions and 57 arrests. - Essex Police recorded zero false positives during those deployments.

The details: - The deployment replaced a more manual approach that depended on officers reviewing CCTV after incidents or spotting people in person on the street. - Department data indicates the 143 positive interventions would not have been feasible under that ad hoc process. - The technology is designed to give officers real-time operational intelligence so they can identify persons of interest and make faster decisions in the field. - The force operated the system at a threshold setting of 55. - The College of Policing guidance discusses an acceptable false positive rate of about 1 in every 1,000 faces scanned during live facial matching deployments. - Corsight AI said the results show facial intelligence built for real-world conditions can help officers act with greater confidence and reduce unnecessary interactions. - Tony Porter OBE QPM LLB, Corsight AI’s chief privacy officer and former UK Surveillance Camera Commissioner, said deployment of live facial recognition must be responsible, transparent and within the legal framework that protects the public.

Between the lines: - The numbers are being used to argue that independently evaluated AI can move from pilot projects to everyday policing. - The zero-false-positive result is especially notable because live facial recognition has faced scrutiny over accuracy, oversight and public trust. - Essex Police’s framing suggests the value of AI is not only identification, but also reducing the administrative burden on frontline officers. - The broader message is that public safety agencies may be more willing to scale AI when human oversight and governance are built into the workflow.

What's next: - Essex Police intends to keep using live facial matching to locate people with outstanding warrants and gather intelligence during active investigations. - Police AI UK is continuing efforts to accelerate responsible AI adoption across UK forces. - The Essex rollout is likely to be used as an example by other agencies weighing whether to expand facial intelligence tools in policing.

The bottom line: - Essex Police is pointing to live facial matching as evidence that AI can deliver faster decisions, more arrests and fewer unnecessary interactions when the technology is independently evaluated and tightly governed. - More information: Corsight AI