Why retail needs biometrics – the cameras aren’t working

By Professor Fraser Sampson, former UK Biometrics & Surveillance Camera Commissioner

According to the Economist shoplifting has octupled in a decade, and now amounts to around 20 million thefts annually.

Is this because we have more shops or fewer scruples? Perhaps, but it’s not confined to shopping. Fare evasion on London’s Tube and bus network has reportedly doubled since 2019 and street crime like phone grabs has made us the snatchers’ capital. That may be a tick for the ‘less scruples’ theory but, while others try to diagnose the underlying causes, let’s stick with the numbers. Think of this: whether they’re supermarkets or corner shops, pop-up vendors or department stores, every one of the venues behind those statistics will almost certainly have some form of security cameras in place. As for the wider picture, with over 70, 000 public-facing cameras, London Underground has one of the most concentrated surveillance systems in the country if not the world, while the square footage of our streets and bus stations and parks covered by CCTV has never been higher. It’s not really going out on a limb then to say that the cameras aren’t working.

When I read recently of the UK’s new ‘one in one out policy’, I expected the article to be about immigration. In fact it was describing how some London shops – despite having CCTV – have decided they must now lock their doors and allow one individual shopper entry only as another leaves. Meanwhile, the British Retail Consortium figures mean the number of retail staff assaulted at work would fill the O2 Arena every 10 days. The figures are stark but the challenge for retail is simple: how to stop nasty people walking in the door and £millions walking out. On any view, the answer can’t be more cameras.

Some retailers have realised that their cameras aren’t working and have instead turned to biometrics. With stores reporting reductions of up to 70% – not just at the start but as a sustained outcome – live facial recognition (LFR) technology is helping retailers take back control of their businesses. The most effective way to stop people doing unwanted things in your shop is to withdraw their invitation. With CCTV cameras that’s easier said than done and it’s very difficult to tell the few people who should be excluded from genuine customers. LFR is perpetrator focused, empowering staff to deny entry to individuals who have previously caused them harm or loss; they can do it quickly and confidently and it’s having amazing results. That’s already stopping crime – the next question is what’s stopping more retailers from using the technology?

Ironically the case against LFR usually starts by challenging its accuracy with inaccurate figures. It’s true that some early police algorithms were poor, but the biometric matching algorithms offered by some providers is over 99.99% – that’s as close to perfect as anyone has ever got. That’s NASA-level accuracy, better than some medical or military procedures and light years away from people staring at CCTV monitors. What about errors and misidentification? Used properly, LFR is a decision support tool, it’s not making the identification itself. Ultimately, it’s helping shopkeepers make their decisions and that’s where the occasional misidentification happens – by human error, not technical. We insist on them being in the loop but “human failsafe” to a robot is an oxymoron.

Some LFR critics argue that it’s too intrusive but CCTV cameras capture the images of everyone coming into, leaving, loitering or even passing by a shop at multiple entry and exit points, car parks, lifts and shopfloor areas – and the operator can keep them for a long time, sharing them with the police and other bodies. Those images can also be analysed by retrospective matching technology for years to come. LFR doesn’t do any of that. If it’s not looking for you it doesn’t even see you let alone take and keep your picture. CCTV is deep sea trawling, LFR is spear fishing.

When making their case against LFR, campaigners invariably wheel out British author George Orwell as their first expert witness against the state. But his world was fictional and, in any event, was based on CCTV cameras. Orwell didn’t predict where we are today. His future had more and more state-controlled cameras and screens not private shopkeepers using technology to provide a safe and secure shopping environment. We the citizens use auditable biometrics routinely for banking, communication, digital security and data protection. The fable of Nineteen Eighty-Four is nothing like the reality of Twenty Twenty-Five. Isn’t it time fiction left the factual debate?

As for its wider application by the state, there’s a lot of barking about the police using LFR but it’s often misdirected – wrong dog, wrong tree. If you don’t trust your police to use new technology, the problem isn’t the technology.

The UK government is supportive of the police using FRT to tackle crime areas, but the technology’s strength for retailers is in preventing the harm before it’s happened. Stopping the 50, 000 daily thefts and 2000 assaults as opposed to expecting the police to do something about them afterwards. LFR is not it’s not plug-and-play CCTV and it’s not fiction. CCTV is photography, LFR is biometrics – two different disciplines. The impact of the latter on retail crime is measurable, verifiable fact while after 50 plus years and millions in public money, the same still cannot be said of CCTV. The available evidence shows that there is an answer to burgeoning retail crime. It isn’t more cameras and can’t be more police. It’s biometrics.

About the author

Fraser Sampson, former UK Biometrics & Surveillance Camera Commissioner, is Professor of Governance and National Security at CENTRIC (Centre for Excellence in Terrorism, Resilience, Intelligence & Organised Crime Research) and a non-executive director at Facewatch.

Article Topics

accuracy  |  biometric matching  |  biometrics  |  criminal ID  |  facial recognition  |  Fraser Sampson  |  London Metropolitan Police  |  real-time biometrics  |  UK