What Is Live Facial Recognition? Met Police Trial

London's Metropolitan Police has started using facial recognition on officers' phones, in a six-month trial that lets police photograph a person and check their face against a database of about one million custody images without asking for consent. The technology, known as Operator Initiated Facial Recognition, or OIFR, began on 17 November and was formally announced on Friday 20 November 2026. It is the latest and most personal form of a technology that has divided opinion in Britain for years, and it raises a simple question that many Londoners are now asking: what exactly is live facial recognition, and how does it work?
Facial recognition is a form of biometric identification. A camera captures an image of a face, software measures the distances and proportions between facial features, and it turns those measurements into a mathematical template. That template is then compared with templates stored in a database. If the software finds a close enough match, it returns a possible identity. If it finds none, the search comes back empty. The process takes seconds, and it can be done from a distance, which is what makes it different from checking a driving licence or asking someone for their name.
There are two main ways police use the technology, and the difference matters. The first is live facial recognition, usually shortened to LFR. In a live deployment, cameras are fixed in a public place, such as a shopping street or a station entrance, and every face that passes is scanned and compared in real time against a watchlist of wanted people. The second is retrospective facial recognition, where footage recorded earlier is searched after an event, for example to trace a suspect's movements after a crime. The Met's new trial is a third variation: operator-initiated recognition. Instead of fixed cameras scanning a crowd, an individual officer takes a photograph of a specific person on a Met-issued mobile phone and runs that single image through the system.
How the Met's new phone trial works
Under the trial, 100 trained officers working in custody suites and on frontline duties carry the OIFR app on official Met phones. When an officer needs to confirm who someone is, for example if a person cannot or will not give their details, or if the officer believes the details given may be false, the officer can take a photograph and upload it into the app. The app then searches the Met's custody image database, which holds around one million photographs taken when people were arrested or detained.
The software behind the trial is NEC's NeoFace algorithm, which has previously been tested by police forces in South Wales, Gwent and Cheshire. If the system finds a match, the officer is shown the possible identity and can then check it against other information, such as police records, before deciding what to do next. A match from the app is not treated as proof of identity on its own. Officers are expected to confirm the result through further checks, and the final decision remains a human one.
The Met says the point of the tool is speed. Confirming someone's identity at the roadside or in a custody suite can currently take a long time, particularly when a person gives no name or gives a false one. A fast identity check can mean a wanted person is detained rather than released, a vulnerable person is identified and helped, or an innocent person is cleared quickly and sent on their way. In the first 24 hours of the pilot, officers uploaded 15 photographs, a figure the Met's Commander Nick John said was higher than he had expected.
Do officers need your consent?
No, consent is not required. That is the part of the trial that has attracted the most attention. Officers must tell a person before taking their photograph, but the person cannot refuse to have the picture taken for this purpose. The Met has said officers will not use force if someone covers their face, and the photograph must be taken on a Met-issued phone, not on an officer's personal device.
There are limits on who can be scanned. The technology will not be used on anyone believed to be under 13, and during the first four weeks of the pilot it will not be used on anyone believed to be under 18. Those age limits reflect the particular sensitivity of collecting biometric data from children and young people, and they are likely to be watched closely by campaigners and by the regulators who oversee police use of the technology.
The legal position is that police use of facial recognition is governed by data protection law, human rights law and the Met's own policies, and the force has said the pilot will be closely monitored, with regular reviews of performance, accuracy and usage data. That monitoring is meant to check both whether the technology works as claimed and whether it is being used fairly, including whether it produces disproportionate results for any group of people.
Why facial recognition is controversial
Supporters of the technology argue that it is simply a faster way of doing something police have always done: recognising people they are looking for. They point out that officers already compare faces with photographs, whether from memory, from a printed image or from CCTV. Facial recognition, in this view, does the same job more quickly and more reliably, and it can help catch dangerous offenders who would otherwise slip away because their identity could not be confirmed in time.
Critics raise three main concerns. The first is accuracy. Facial recognition systems have been shown in independent testing to be less accurate for some groups, including women and people with darker skin, raising the risk that innocent people could be wrongly flagged as suspects. A wrong match can lead to a wrongful stop, a wrongful arrest or a damaging encounter with the police, even if the error is corrected later. The second concern is privacy. A face, unlike a password, cannot be changed, and campaigners argue that scanning faces in public without consent amounts to a form of mass surveillance, even when the scan is of one person at a time rather than a whole crowd. The third concern is scope creep: a tool introduced for a narrow purpose, such as identifying wanted offenders, may gradually be used for wider purposes once it is in place.
These concerns are not new. An inspection by His Majesty's Inspectorate of Constabulary and Fire and Rescue Services has previously examined the risk of wrongful arrest from facial recognition, and the Home Office has reviewed the privacy implications of the technology. The Met's trial is therefore taking place in a context where the safeguards, and whether they are strong enough, are already a matter of public debate.
What happens to your photograph?
One of the most common questions about the trial is what happens to a photograph once it has been taken. The Met has said the images are compared against its custody database, and that the pilot will be reviewed regularly. What the force has not yet published in full is a detailed account of how long non-matching images are kept, who can access them, and how a person can find out whether their image was searched. Those details matter, because data protection law gives people rights over their personal data, including biometric data, which is treated as a particularly sensitive category.
If the trial is extended beyond six months, or rolled out more widely, those questions are likely to become the centre of the argument. A wider programme would also have to decide whether the under-13 exclusion remains, how officers are trained to use the app, and what happens when the system produces a match that turns out to be wrong.
What happens next?
The pilot runs for six months from 17 November, which means it is due to end in May 2027. During that time the Met will collect data on how often the tool is used, how often it returns a match, how accurate those matches are, and whether its use differs across different groups of people. At the end of the trial the force will decide whether to seek a wider rollout. That decision will be made against a backdrop of court cases, parliamentary scrutiny and public debate about where the line should be drawn between effective policing and personal privacy.
For now, the practical position for Londoners is straightforward. An officer may photograph you without your consent if they are taking part in the trial and need to confirm your identity, but they must tell you first, they must use an official phone, and they must not use force if you cover your face. The technology is new, the trial is limited, and the debate about whether it should become a permanent part of British policing is only just beginning.
Sources: Metropolitan Police | Biometric Update