How automatic number plate recognition works
Automatic number plate recognition, universally known as ANPR, reads a vehicle's registration plate as it passes a camera and checks it instantly against police databases, including the Police National Computer. A match, referred to as a hit, can indicate a stolen vehicle, an untaxed or uninsured one, a vehicle linked to a person wanted for arrest, or one otherwise flagged of interest to an investigation. Officers can then act on that alert under section 163 of the Road Traffic Act 1988, which gives police the power to stop a vehicle. ANPR is not new artificial intelligence in the sense that facial recognition or predictive analytics are, but the underlying character recognition and the systems built on top of it, including automated pattern analysis of vehicle movements over time, increasingly rely on machine learning.
It is, in scale, the most heavily used automated identification system in British policing. Cameras across England, Wales, Scotland and Northern Ireland submit tens of millions of plate reads to the National ANPR Data Centre at Hendon every day, a volume that dwarfs any other form of automated police surveillance in the country, including facial recognition, by a considerable margin.
The national data centre and retention
Every police force's ANPR reads are pooled at the National ANPR Data Centre, alongside the Police National Computer. Published guidance from the National Police Chiefs' Council and individual forces states that this data is retained centrally for around twelve months, though earlier documentation and some campaign groups have cited a two year retention period for certain categories of record, and the precise position has shifted over time as national standards have been updated. Access to the data after collection is tiered: most authorised staff can only query records up to 90 days old, with access to older data restricted to senior officer authorisation and, in many forces, limited to serious, major or counter terrorism investigations.
The practical effect of this system is that the location and time of a very large proportion of vehicle journeys made on major roads in the United Kingdom is captured and held centrally, whether or not the vehicle or its occupants are suspected of anything. National ANPR Standards for Policing and Law Enforcement, alongside separate audit standards, are intended to constrain who can search that data and for what purpose, but the sheer scale of what is collected, applying by default to the innocent and the suspected alike, is what has made ANPR a recurring subject of scrutiny from privacy campaigners even though it attracts far less public attention than facial recognition.
Where the technology is heading
The Home Office issued a preliminary market engagement notice in October 2025 for a new National ANPR Service Application, seeking a replacement or substantial upgrade to the existing national system, with a stated scope covering live time alerting, control room deployment, search and reporting, and data preservation and audit functionality. That procurement signals that the ageing national infrastructure behind ANPR is being renewed, which matters because decisions taken during a procurement of this kind, about what data the new system retains, for how long, and who can query it, will shape ANPR's privacy footprint for years to come with comparatively little public debate, since procurement decisions of this sort rarely receive the same scrutiny as a new frontline technology being announced.
Separately, ANPR has an increasingly close relative in the private sector. Camera systems such as Flock Safety, widely deployed by police departments and by private homeowners associations in the United States, perform a similar function using commercially owned cameras and, in some configurations, share vehicle sighting data with law enforcement outside the kind of standardised, centrally audited framework that governs the UK's national system. Coverage on this site follows both the evolution of the UK's own national ANPR infrastructure and the growth of these adjacent, sometimes less regulated, commercial equivalents.
What is not settled
There is no single, current and authoritative published figure for national ANPR retention that all sources agree on, itself a small illustration of a wider transparency gap: a system that captures more data on more people than any other form of police technology in the country operates under standards that are not always consistently or clearly stated in public facing material. Nor is there a great deal of independent, judicially tested case law comparable to the Bridges ruling on facial recognition that has examined whether blanket retention of ANPR data on people never suspected of anything is proportionate. As the national system is renewed through the Home Office's current procurement, whether that renewal is accompanied by clearer public retention standards, or simply preserves existing practice inside newer technology, remains to be seen.
Follow the coverage
PoliceAI News tracks every anpr story from across the English-speaking world as it breaks: new deployments, court rulings, parliamentary debates, academic research, and community responses. The feed refreshes every 30 minutes.
View Live ANPR StoriesYou can also browse every anpr story in the feed, or explore every subject PoliceAI News covers.