Recent years have witnessed fierce debates on the merits and flaws of the use of Live Facial Recognition (LFR) technology by British police, a pendulum swinging from its alleged crime prevention results to its protested effects on people’s right to privacy.
LFR was first deployed by the Metropolitan Police Service (MET) at the 2015 Notting Hill Carnival and is currently used by almost all major police forces in the UK. Its deployment is on the rise: the MET used it 32 times between 2020 and 2023, and 231 times in 2025.
LFR relies on a software-based facial matching tool which can be applied to videos or image feeds like CCTVs or custody databases without the need for installation. The software first detects a face; it then scans the biometrics of the ‘subject’ and converts them into a unique identifier – the ‘faceprint’; lastly, this digital signature is cross-referenced with other faceprints uploaded by the police on the ‘watchlist’ to find a match. The police use a static camera to record and stream footage in real time to the software. Upon finding a match, an alert is sent to a control room and to on-the-ground officers, both of whom have discretion to act on that alert by approaching the subject. Current watchlists span over 15,000 people, with deployments lasting usually around 6 hours with an average of 15,000 faces scanned per session.
Legal status
Currently, there is no UK legislation specifically regulating the use of LFR. Nonetheless, the technology has faced legal challenges due to its impact on the rights to privacy, freedom of expression and freedom of assembly. These rights, afforded to us by the European Convention of Human Rights and the 1998 Human Rights Act, are qualified rights, so an interference with them can only be justified if it is:
- For a legitimate purpose,
- In accordance with the law (Art.8) or prescribed by law (Art.10, 11), and
- Necessary in a democratic society.
In the recent High Court case of The King (Shaun Thompson and Silkie Carlo) v The Commissioner of Police of the Metropolis and The Equality and Human Rights Commission (Intervener) the claimants contended that the 2024 MET LFR policy, which except for some changes is still in operation, did not have the quality of law as it left too much discretion to police officers as to where, why and against whom LFR may be used.
The MET Policy, at the time of the challenged policy and now, presents three relatively broad scenarios to determine when and where LFR can be deployed, that is for:
- Crime hotspots and missing-person hotspots,
- Protective security operations, including major events and critical national infrastructure,
- Deployments based on specific intelligence indicating that a sought individual is likely to be present at a particular location.
The list of people who could be placed on a watchlist includes three main categories:
- Those wanted on a court warrant or subject to certain court orders,
- Vulnerable and missing persons,
- Those suspected of having committed, committing or about to commit, a crime.
Before authorising a deployment, an Authorising Officer must consider whether it would be a proportionate means of achieving the MET’s policing objectives, in light of the impact of deployment on the rights and freedoms of members of the public.
The Divisional Court unanimously dismissed the claim. The judges considered that the MET policy’s criteria regarding watchlist, deployment, operation and oversight were detailed and objective enough to have the ‘quality of law’. In particular, the judges cited the mandatory proportionality assessment which Authorising Officers need to carry out before deciding whether to deploy the technology. The claimants have stated their intention to appeal.
The limits of human rights
The Thompson case epitomises the limitations of human rights law to criticise or explain the operation of LFR. To talk about LFR in the language of human rights means that the primary question is one about the legitimacy of state interference in our lives and privacy. The provision and protection of human rights is ultimately always dependent on the state, hence why the human rights mechanism does not seek to question state interference, only to regulate it. This results in the police imposing superficial human-rights compliant safeguards, and the courts accepting them. In fact, the police implemented to the letter the suggestions of the Court of Appeal in the 2020 R v Bridges case, in which the South Wales Police’ LFR policy was deemed to be not ‘in accordance with the law’ as it was not detailed enough.
However, these checks and controls overlook the heart of the matter, in as much as the language of human rights fails to ask the right question. What is shrouded beneath the rights-compliance discourse is an almost bipartisan assumption that LFR is a ‘highly effective and efficient policing tool’, a ‘precision crime-fighting tactic’. The media, legal and political discourse has so far unquestionably accepted this narrative, one which relies on an idea of crime as something static and objective, which can be simplified to measurable data primed for quick-fix technical answers.
In light of the massive overextension of policing responsibility following the austerity cuts of 2010-2019 to statutory, social and mental health services as well as to the police itself, LFR was an alluring prospect which could help the police do more with less. But its perceived objectivity belies a belief that crime can be reduced not by addressing its root causes or social conditions, but rather by acting directly on the data generated by it. However, for data to be acted upon, it needs to be purified, abstracted from its context. Data is by default de-contextualised, and as such acting purely on data ultimately conceals and de-politicises the complex structural issues of inequality which fester within the criminal justice system.
Fixing the justice system
The criminal justice system, including the penal system and the police, is unsurprisingly riven by long-standing rifts along lines of ethnicity, gender, nationality and class. Black and mixed-race people, especially children, are disproportionately represented in all aspects of the penal system, i.e. stop and searches, convictions and prison population. Social characteristics also shed light on the picture: people who never had a job, were taken into care as children or have attempted suicide at some point in their lives form a much greater proportion of the prison population than they do of the general population. The appeal of LFR is that, rather than addressing these long-term issues of structural inequality, it proposes a seemingly innocuous and technical fix.
Conveniently, this chimes in with some politicians’ craven promises to be ‘tough on crime’, which result in ill-thought vote-grabbing proposals such as limiting the right to jury trial. However, the extent to which LFR, and by extension the criminal justice system itself is equipped to deal with society’s ills is grossly misconceived. A structural example is the youth justice system. England and Wales had an average daily custody population of 420 children in 2024, with around 13,000 being sentenced in that year. Currently, 6 in 10 children released from custody reoffend within a year, and 80% of children cautioned or sentenced are neurodivergent or have SEND. Children in care are 10 times more likely than other children to ever receive a caution or conviction.
These shortcomings are systemic. The uniquely low age of criminal responsibility in England and Wales (10, compared to the average 14.5 in OECD countries), together with the chronic underfunding of key agencies means that issues that should be addressed early on by social services or child psychologists are delegated to the courts, which lack funding for truly ‘child-first’ procedures.
Justice’s blind arms have in recent years extended to embrace more and more political and societal issues, the brief mention of which will suffice to bring to mind the failure of the criminal system in actually ameliorating these problems. The shoplifting epidemic, the mental health crisis, the environmental, pro-Palestine and far-right protests, are all examples of deeper, structural issues which cannot be solved by purely legal or technical solutions.
How can we help
We know the system has its flaws — but we know how to work within them for you. We regularly challenge the admissibility of evidence obtained by the police and fight for our clients’ rights in Court. If you have been identified by LFR or have been accused of committing a crime, get in touch with our experienced Crime team.
You can contact us for a free, no obligation initial discussion on 020 3440 8000 or email enquiries@tvedwards.com.
For an extended version of the analysis behind this article, see: Adalberto Dionisi, ‘What Live Facial Recognition does not see: Human Rights, Discipline and Governmentality’ (2026) 11(1) LSE Law Review
Disclaimer: The information on the TV Edwards website is for general information only and reflects the position at the date of publication.
