For shoppers frequenting the Sainsbury’s supermarket in Sydenham, south London, over the festive period, there was the usual seasonal battle to pile trolleys high with Christmas treats and manoeuvre them through crowds of families doing the same thing.
What few of these people realised, however, was that every customer making their way to the checkout was being filmed, with their facial features scanned to check whether they matched a database of known shoplifters in the UK.
In September, Sainsbury’s announced it was trialling facial recognition technology in the store in an effort to cut both theft and abuse directed at staff, a move it says could be rolled out across its estate nationwide.
Sainsbury’s is far from alone. Facial recognition technology is increasingly being deployed by British retailers, driven by advances in AI and the falling cost of smart cameras.
For years, CCTV cameras were largely “dumb” devices, incapable of understanding what they were capturing, relying instead on security staff to monitor live feeds or review footage after an incident.
But for the last decade or so, machine learning algorithms trained to recognise specific people, objects, or behaviours have been enabling cameras to detect changes in real time and flag them automatically.
The technology itself is relatively straightforward. Cameras capture images of everyone entering the store, with software locating and isolating faces within the frame. Unique facial landmarks, known as nodal points, such as the distance between the eyes, the depth of eye sockets, cheekbone shape, and jawline contour, are measured and converted into a numerical code, often referred to as a faceprint or facial template.
A number of tech companies have been quick to spot an opportunity and have been rolling out services offering to detect those placed on store blacklists for previous offences of shoplifting or anti-social behaviour.
According to the UK Office of National Statistics, the number of people who have been caught and convicted of shop theft has risen sharply since the pandemic, increasing 20% in 2025 compared with the previous year.
Sainsbury’s says that by introducing the technology it is helping to protect its staff and its stock from rising levels of theft and anti-social behaviour.
“After 17 years in retail, including the last year managing our Sydenham store, it’s easy to feel like you’ve seen it all. But I have never witnessed theft, anti-social, and threatening behaviour at the level we are seeing now,” says store manager Kashif Rasoo. “These actions impact my team and our customers, who all too often witness, or even worse, become the target of threatening incidents and theft. This has to stop. Facial recognition technology adds an extra layer of security that gives us the confidence to be one step ahead of people committing criminal acts in our store.”
The technology used by Sainsbury’s is provided by Facewatch, which says it also works with a range of other High Street retailers, including Budgens, Sports Direct, B&M Bargains, and Southern Co-op.
According to Facewatch CEO Nick Fisher, if store staff identify that a customer has committed a crime, they can log into the system and add that individual to a database. Should the same person return, the system sends an alert to staff, which is checked by two algorithms and, where necessary, a human verifier.
Although the company acknowledges that mistakes have occurred, it claims an accuracy rate of 99.98% and says the technology has led to retail crime falling by up to 70% in some stores.
Facewatch told IoT Insider that the company sent out a total of 54,312 in December, a monthly record. And in the week running up to Christmas, traditionally the busiest shopping period of the year, it sent its highest ever weekly total – 14,885 an increase of 3,000 compared with the company’s previous highest weekly total.
Retail is just one sector adopting AI-enabled surveillance. Governments and law enforcement agencies use the technology daily, while millions rely on it to unlock their phones.
Private firms are rapidly finding new use cases, from theme parks speeding up entry queues to doctors diagnosing rare genetic disorders or apps detecting those having strokes, construction companies enforcing safety compliance, reuniting missing children with their families in India, and estimating the age of customers buying alcohol.
Critics, however, warn that the technology can be overly intrusive, risks infringing civil liberties, and may create data security vulnerabilities. Others argue that facial recognition algorithms disproportionately misidentify women and people of colour.
In China, a global leader in facial recognition, authorities were forced to apologise after using the technology to publicly shame people for minor offences, including wearing pyjamas in public, lying on benches deemed “uncivilised”, and handing out advertising flyers.
Such concerns have led at least 14 cities, mostly in the USA, to ban the use of facial recognition by local agencies, including police forces and transport authorities.
In the UK, campaign groups such as Liberty and Big Brother Watch have been campaigning to introduce similar measures, demanding safeguards on the technology and to stop the use of facial recognition in public spaces. Their case has been taken up by MPs on both sides of the House and was debated in parliament in 2024.
“Facial recognition surveillance is out of control,” says Silkie Carlo, Director of Big Brother Watch. “We are hurtling towards an authoritarian surveillance state that would make Orwell roll in his grave.”
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by visiting our LinkedIn page.
