The Metropolitan Police has made more than 100 arrests in Croydon during the first three months of a new Live Facial Recognition (LFR) pilot, which for the first time in London uses cameras mounted on existing street infrastructure rather than police vans.
The pilot, which began in October 2025, places LFR cameras on lamp posts and other fixed structures along Croydon’s high street. According to new figures released by the Met, the approach has significantly increased efficiency, with an arrest made on average every 34 minutes during deployments.
A third of those arrested were wanted for violence against women and girls (VAWG) offences, including strangulation and sexual assault. Other arrests were linked to burglary, possession of offensive weapons, and recall to prison. The Met says crime in Fairfield Ward has fallen by 12% since the pilot began, including reductions in retail crime, violent crime and sexual offences.
Lindsey Chiswick, the Met’s national lead for LFR, said the Croydon pilot demonstrates the technology’s growing value in tackling serious crime. “More than 1,700 dangerous offenders have been taken off London’s streets since the start of 2024, including those wanted for rape and child abuse,” she said. “The amount of arrests we have made in just 13 deployments shows the technology is already making an impact and helping to make Croydon safer.”
The Met says LFR now locates wanted individuals more than 50% faster than van based deployments and far more quickly than traditional policing methods such as door to door enquiries. Three quarters of those arrested during the pilot live in Croydon, which the force says reflects its targeted, intelligence led approach.
Croydon’s Executive Mayor, Jason Perry, welcomed the results. “The pilot has led to a significant number of arrests, including individuals suspected of serious offences and violence against women and girls,” he said. “This pioneering technology is helping to make our streets safer.”
How the pilot works
Cameras are mounted on existing street furniture at two locations on Croydon High Street. Feeds are monitored remotely, freeing up police vans for other operations. Cameras are only activated when officers are present.
Each deployment uses a bespoke watchlist created within the previous 24 hours and deleted immediately afterwards. Specialist LFR officers and neighbourhood teams remain on site to engage with the public and respond to alerts.
The Met says Croydon was chosen due to its status as a crime hotspot and the success of previous LFR deployments in the area. Including van-mounted deployments, since the start of 2024, LFR operations in Croydon have resulted in 249 arrests, with 193 suspects charged or cautioned.
There are currently no plans to expand the pilot to other boroughs.
Legal evaluation
The pilot will now undergo formal evaluation to assess its effectiveness, while engagement sessions with residents and councillors continue to address privacy safeguards and operational transparency.
The fixed camera pilot represents a shift in how LFR is deployed, how data is processed, and how the Met frames its compliance narrative. For the first time in London, LFR cameras were mounted on existing street furniture rather than police vans.
Fixed cameras create a more persistent surveillance presence, even if activated only during deployments. This raises questions about whether the pilot edges closer to “overt public space surveillance” which may bring it into conflict with the Protection of Freedoms Act 2012 (PoFA) and the Surveillance Camera Code of Practice. The Met will need to demonstrate that fixed point deployments are no more intrusive than van based operations and that the shift is justified by operational gains (e.g., faster arrests, reduced resource burden).
LFR also engages multiple strands of data protection law:
• Biometric data is “special category data” under UK GDPR and requires a lawful basis under both Article 6 and Article 9.
• Automated processing that identifies individuals engages Article 22 safeguards.
• Article 8 ECHR (right to private life) requires clear justification, safeguards, and oversight.
The Met asserts that watchlists are created within 24 hours of deployment and are deleted immediately afterwards and that cameras are only activated when officers are present.
These safeguards do not remove the need for a robust Data Protection Impact Assessment (DPIA), clear articulation of the lawful basis (usually law enforcement processing under Part 3 of the Data Protection Act 2018) and demonstrable minimisation of false positives and demographic bias. The Met’s claim of 85% public support does not, in itself, satisfy legal proportionality tests.
FOI implications: transparency vs operational sensitivity
The pilot raises several FOI relevant issues that requesters will inevitably probe:
a. Watchlist composition
The Met states that each deployment uses a “bespoke, intelligence led watchlist”. Under FOIA, requesters may seek criteria for inclusion, categories of offences and sources of intelligence. The Met will likely rely on section 31 (law enforcement) to withhold specifics, but high level criteria may be disclosable.
b. Data Protection Impact Assessments (DPIAs), governance documents and audits
These are routinely requested in LFR contexts. The Met has historically disclosed redacted DPIAs, but exemptions under s.31 (law enforcement), s.38 (health and safety) and s.24 (national security) are often applied to operational detail.
Given the shift to fixed infrastructure, requesters may also seek location maps of camera sites. technical specifications, accuracy and false match rates.
c. Statistical transparency
The Met has voluntarily published arrest numbers, charge/caution outcomes, and efficiency metrics. Under FOIA, these are generally disclosable unless they risk identifying individuals. The fact that 193 of 249 LFR linked arrests in Croydon since 2024 have resulted in charges or cautions may be used to support proportionality arguments but FOI requesters may seek more granular breakdowns.
Governance and oversight
The Met will need to demonstrate that the use of fixed infrastructure in the Croydon pilot meets the standards set by the Court of Appeal’s decision in R (Bridges) v South Wales Police [2020] EWCA Civ 1058 (https://www.judiciary.uk/judgments/r-bridges-v-cc-south-wales/).
This considered the legality, proportionality, and transparency surrounding the use of facial recognition technology in public spaces and remains the leading authority on LFR. It requires:
• Clear and accessible policies.
• Sufficiently detailed guidance on watchlist creation.
• Non discriminatory deployment.
• Adequate data protection impact assessments.
The Met will need to demonstrate that the Croydon pilot meets these standards, particularly given the novel use of fixed infrastructure. The Met says that it is aware of these obligations, but the forthcoming evaluation will be crucial. If the pilot expands, renewed scrutiny from civil liberties groups and the Information Commissioner is very likely.
Local authority involvement and shared accountability
Croydon Council’s active support for the pilot introduces an additional governance dimension. As councils are subject to the FOIA and the Environment Information Regulations, meaning requesters may seek correspondence, agreements, and risk assessments held by the local authority. Councils must ensure their involvement aligns with PoFA and the Surveillance Camera Code, even if the Met is the data controller for biometric processing.
What to watch next
The Met says there are no current plans to extend the pilot. The evaluation phase will scrutinise:
• DPIA adequacy
• False match rates
• Equality impact assessments
• Watchlist governance
• Public engagement records
• Retention and deletion logs
The legal framework for LFR may face renewed calls for statutory reform, something both the ICO and parliamentary committees have previously urged.
