Ghost plates are outrunning the cameras built to catch them

At the Dartford Crossing, enforcement cameras are now logging more than a thousand number plates a day that do not match any vehicle actually on the road.

On one M25 site, that daily count has risen from roughly 1,000 to nearly 2,000 over the past two years, according to automatic plate readers filmed for BBC Panorama’s investigation into so called ghost plates. Insurance Business UK

Parliament’s own inquiry into illegal registration plates puts the scale higher still. It estimates that as many as one in fifteen vehicles on UK roads may be carrying a plate altered to defeat that detection, against 34,455 licensed plate suppliers and only 14 DVLA prosecutions in the first ten months of 2025. APPGTS, Ghosts on the Road

Here is the detail worth sitting with. Every lawful policyholder already pays into the Motor Insurers’ Bureau levy, which exists to compensate people hurt by drivers who were never properly insured in the first place. That levy was built for a world where an untraceable driver was a rare failure of the system.

A plate designed specifically to defeat the cameras is not a failure of the system. It is a product, bought from a supplier who is one of tens of thousands with almost no realistic chance of being caught.

Detection has improved faster than enforcement, and claims teams are relying on evidence that knows it.

A theatre can check every ticket carefully at the door. It can still sell the same seat twice if nobody ever counts how many tickets left the printer. The camera at the door is not the weak point. The uncounted supply behind it is.

This matters to claims handling in a way that goes beyond a fraud headline. ANPR data increasingly stands in for a driver’s own account of where they were and when, in everything from low speed collision disputes to hit and run investigations.

If a meaningful share of plates on the road are not what they claim to be, that evidence base is less dependable than claims teams currently treat it as being. Nobody has published what share of disputed claims already turn on a plate read that cannot be fully trusted.

None of this is free for the people who did nothing wrong. A claims handler who treats an ANPR hit as settled fact, rather than as a probability, is one misread cloned plate away from treating a genuine policyholder as the liar in their own claim.

Consumer Duty asks firms to deliver fair outcomes, not merely efficient processes. An efficient ANPR network built on an unreliable plate population is efficient at the wrong thing.

The honest fix sits upstream of insurance altogether, with plate suppliers and the prosecutions meant to deter them. Fourteen prosecutions against tens of thousands of suppliers is not a deterrent, it is a rounding error.

Insurers are not bystanders here, even so. They are the industry paying out when the evidence turns out to rest on a plate that was never real, and they are the ones best placed to say publicly how often that already happens.