Boardmasters ran across the first weekend of August in Newquay, with something like 50,000 people over three days and Fatboy Slim on the bill. Devon and Cornwall Police parked a van in Killacourt, a public square near the seafront, fitted it with a biometric camera, staffed it with six officers, and ran live facial recognition against a watchlist for the first time at that event. The watchlist held thousands of entries: people wanted for offences, people subject to sexual harm prevention orders, people suspected of dealing.
Almost every write-up framed this as festival security. Look at where the van was and that framing falls apart.
The scan happened outside the ticket
Killacourt is a public green in the middle of Newquay. It is not the festival site, it is not behind a wristband check, and walking through it is not an act of consent to anything. The people whose faces were matched against a database that weekend included festivalgoers, and they also included residents walking to the shops, families on holiday, and teenagers who could not afford a ticket.
That distinction matters more than the technology does. A search at a festival gate is a condition of entry, published in advance, and avoidable by not going. A camera in a town square is none of those things. The event supplied the reason and the crowd, and the deployment covered everyone the crowd was mixed in with.
This is the pattern worth watching as the technology spreads, and it is spreading. Vans have since been deployed in Leeds. A large gathering makes an unarguable case for a trial, and the trial is conducted on the surrounding public space rather than the gathering.
The case for it is real, and it is narrow
It would be dishonest to pretend the police case is empty. Sexual offending at festivals is a genuine and underreported problem. Organised dealing at large events is real, and the people running it are not the people taking the risk. If a camera identifies someone subject to a sexual harm prevention order standing in a crowd of teenagers, that is a good outcome and most people would want it.
The question is not whether the aim is legitimate. It is whether the method stays attached to the aim once the equipment exists and the vans are already bought.
Watchlists are the pressure point. Boardmasters' list ran to thousands and mixed categories that carry very different weights: convicted offenders, people merely suspected, and people subject to preventative orders. Those are three different legal positions being enforced by one camera at one threshold. A system that produces defensible results against the first category will produce far more arguable ones against the third.
Nightlife has been the testing ground before
Clubs and festivals are where surveillance gets normalised, and this is not new. ID scanners on doors that retain data long after the night ends. Cashless wristbands that turn every drink into a record. Airport-style search as a routine condition of entry. Each arrived as a response to a specific harm, each stayed, and none of them were subsequently removed when the harm receded.
The reason nightlife goes first is structural rather than sinister. The crowds are young, the political capital in objecting is low, and the venue carries a licence that can be threatened. A promoter asked to accept a new condition is not in a strong position to refuse, because the licence is the business.
So the sequence goes: introduced at events, accepted at events, deployed in the town the event happens in, then deployed in the town.
What an organizer can actually do
Very little about a police deployment on public land, and more than nothing about the parts inside the fence.
- Ask what is being deployed and where, in writing, before you agree a site. If a force plans a van near your event, you will not be able to stop it, and you can find out in time to tell your audience.
- Publish it. People have a right to know that attending means being scanned, and telling them costs you nothing except the discomfort of saying it.
- Know what your own systems keep. Door scanners, ID checks and ticketing platforms all accumulate identity data, and most organizers have never asked how long theirs is held or who can request it.
- Separate what you need from what you collect. A headcount does not require a name. A capacity limit does not require an address.
That last point is the one we think about at Dizko, because we build in this space. Our own door scanning verifies that a ticket is valid and has not been used twice, which needs a ticket reference and not a face. The design question for anyone building event tooling is whether a feature needs identity or merely needs a count, and those get conflated constantly because identity is easier to store than it is to justify.
The thing nobody measured
The reporting so far has covered the deployment and the objections. What is missing is the outcome. How many matches were made across that weekend, how many were true, how many people were stopped, and how many of those stops led anywhere.
Without those numbers the debate is two sets of assertions. Forces describe the capability and campaigners describe the risk, and the public is asked to weigh a benefit that has never been quantified against a cost that has never been measured. Any force running this at a public event should be publishing the match rate and the false-positive rate afterwards, in the same way it publishes stop-and-search figures.
Until that happens, the honest position on facial recognition at festivals is that we have deployed it faster than we have learned to evaluate it. The next Boardmasters will be easier to justify than this one was, and nobody will have to make the argument again.