This essay will argue that while live facial recognition (LFR) technology is not necessarily incompatible per se with the UK’s human rights framework, its current deployment by police forces is practically incompatible with both Article 8 of the European Convention on Human Rights (ECHR) and domestic equality protections. The decision of the Court of Appeal in R (Bridges) v Chief Constable of South Wales Police unequivocally demonstrated that the existing legal framework is deficient, affording police excessive discretion and failing to provide the specific, foreseeable, and robust safeguards necessary to legitimise such a significant intrusion into fundamental rights. This essay will critically analyse the interference LFR poses to the right to privacy, evaluating the justification test under Article 8(2) and highlighting the legal lacuna identified in Bridges. It will then examine the demonstrable risk of discrimination posed by algorithmic bias, linking this to the Public Sector Equality Duty (PSED). Ultimately, it will conclude that in the absence of bespoke primary legislation, the use of LFR operates beyond a clear and foreseeable legal basis, rendering it incompatible with the rights it purports to respect.
Live Facial Recognition and the Right to Privacy under Article 8
The right to respect for private and family life, enshrined in Article 8 ECHR and incorporated into UK law by the Human Rights Act 1998, is a cornerstone of democratic society. LFR technology, which automates the identification of individuals in real-time against a watchlist, represents a novel and profound challenge to this right. Its compatibility hinges on a two-stage analysis: first, whether its use constitutes an interference with a person’s ‘private life’, and second, if an interference is established, whether it can be justified under the strict criteria of Article 8(2).
Establishing an Interference with Private Life
The concept of ‘private life’ under Article 8 is broad and not confined to a domestic or secluded sphere. The European Court of Human Rights has long recognised that an expectation of privacy can exist even in public places (Von Hannover v Germany (No. 2), 2012). However, LFR technology goes far beyond conventional CCTV surveillance. While CCTV records individuals, LFR actively processes their biometric data – the unique characteristics of their face – to identify them. This systematic capture, processing, and comparison of personal data for every person who passes before the camera constitutes a far more significant intrusion.
The landmark case concerning biometric data is S and Marper v United Kingdom (2008), where the Grand Chamber of the ECHR held that the indefinite retention of innocent individuals’ DNA profiles and fingerprints was a disproportionate interference with their Article 8 rights. The Court emphasised the uniquely personal nature of such data and the state’s obligation to justify its retention. By analogy, facial images processed as biometric data are similarly sensitive. As Professor Lorna Woods argues, facial recognition changes the nature of public space from a place of relative anonymity to one of perpetual identification, creating a ‘chilling effect’ on freedom of expression and assembly (Woods and Dowdney, 2019). Individuals may be deterred from attending protests, meetings, or expressing dissent if they know their presence is being automatically logged and identified by the state.
This reasoning was central to the Court of Appeal’s judgment in R (Bridges) v Chief Constable of South Wales Police [2020] EWCA Civ 1058. The claimant, a civil liberties activist, challenged the use of LFR by South Wales Police (SWP). The court robustly rejected the initial High Court finding that LFR was analogous to a police officer looking for a face in a crowd. Instead, it held that the processing of biometric data from every member of the public passing the camera, not just those flagged as a match, amounted to an interference with the Article 8 rights of all those individuals. The Court concluded that "the Article 8 rights of the appellant were violated" in this respect (Bridges, [67]). This finding confirms that the dragnet nature of LFR, capturing and processing biometric information indiscriminately, is the primary source of the interference with privacy, shifting the debate definitively towards the question of justification.
The Justification Test under Article 8(2)
For an interference with Article 8 rights to be lawful, it must be ‘in accordance with the law’, pursue one or more legitimate aims, and be ‘necessary in a democratic society’. While the aims of preventing crime and ensuring public safety are undoubtedly legitimate, the use of LFR by UK police forces, as analysed in Bridges, fails comprehensively on the first and arguably the third limb of this test.
#### ‘In Accordance with the Law’
The ‘in accordance with the law’ requirement is not merely a formality; it demands that the legal basis for the interference is accessible, foreseeable, and provides adequate safeguards against arbitrary use (Malone v United Kingdom, 1984). The law must be sufficiently clear to give citizens an adequate indication of the circumstances in which the authorities are empowered to resort to such measures. It was on this ground that the legal framework for LFR collapsed under scrutiny.
The SWP in Bridges attempted to justify its LFR deployment using a patchwork of existing laws: its common law powers to prevent and detect crime, the Data Protection Act 2018 (DPA 2018), and internal force policies. The Court of Appeal found this framework to be wholly inadequate. It held that for a measure as intrusive as LFR, the law must define the scope of the discretion conferred on the competent authorities and the manner of its exercise with sufficient clarity. The existing framework failed this test spectacularly. As the Court noted, "there is no clear guidance on where AFR [Automated Facial Recognition] can be used and who can be put on a watchlist" (Bridges, [96]).
The discretion granted to individual police officers was deemed "too broad" (Bridges, [97]). The internal SWP policies did not have the force of law and did not adequately constrain decisions about deployment locations (e.g., political protests, transport hubs, shopping centres) or the criteria for placing individuals on a watchlist. Could a watchlist include not just those wanted for arrest, but also witnesses, missing persons, or individuals deemed ‘of interest’ without any link to serious crime? The legal framework provided no clear answer, meaning its application was not foreseeable by the public. This lack of a specific, statutory basis led the Court to conclude that the use of LFR was not in accordance with the law. This finding goes to the heart of the essay’s question: without a proper legal foundation, LFR’s use is, by definition, incompatible with Article 8.
#### ‘Necessary in a Democratic Society’
The necessity test requires that the interference corresponds to a ‘pressing social need’ and is proportionate to the legitimate aim pursued. This involves balancing the severity of the rights infringement against the benefits gained. Although the Court of Appeal did not need to make a conclusive finding on proportionality, having found the scheme unlawful on other grounds, the judgment raised significant doubts.
The Information Commissioner, intervening in Bridges, argued that the force had not proved LFR to be strictly necessary, as opposed to merely useful (ICO, 2019). The proportionality of LFR is highly questionable given the lack of robust evidence of its effectiveness in preventing serious crime, weighed against the mass infringement of privacy it entails. Critics point to the high rate of false alerts in deployments by both the SWP and the Metropolitan Police, which risks undermining public trust and diverting police resources (Big Brother Watch, 2018).
Moreover, the availability of less intrusive means must be considered. Traditional, intelligence-led policing, while more resource-intensive, does not involve the indiscriminate biometric scanning of the entire population in a given area. Therefore, it is strongly arguable that the deployment of LFR as a general policing tool cannot be considered a proportionate response, and is not ‘necessary in a democratic society’. The pervasive surveillance it enables creates a societal shift that is disproportionate to its unproven benefits.
Live Facial Recognition and Equality Protections
Beyond privacy, a critical axis of incompatibility relates to equality. The concern is that LFR technology does not affect all citizens equally and may systematically discriminate against certain demographic groups, particularly along the lines of race and gender. This engages both the Public Sector Equality Duty under domestic law and the prohibition on discrimination under Article 14 ECHR.
The Risk of Algorithmic Bias
The algorithms at the heart of LFR systems are ‘trained’ on vast datasets of faces. If these datasets are not demographically representative—for instance, if they are predominantly composed of white males—the algorithm may become less accurate at identifying individuals from under-represented groups. Extensive research, most notably by the US National Institute of Standards and Technology (NIST), has consistently confirmed this phenomenon. A major NIST study found that across dozens of LFR algorithms, false positive rates were often 10 to 100 times higher for Black and Asian individuals compared to white individuals (NIST, 2019). Similarly, error rates were generally higher for women than for men.
This is not a theoretical problem. In the UK context, an independent review of the Metropolitan Police’s LFR trials found that the system’s effectiveness was questionable and raised concerns about bias (Fussey and Murray, 2019). A higher rate of false positives for ethnic minorities and women translates into a greater likelihood of these individuals being wrongly stopped, questioned, and subjected to police intervention. This not only constitutes a distressing and humiliating experience but also exacerbates existing tensions between police and minority communities, undermining the principle of policing by consent. The discriminatory application of technology thereby reinforces and amplifies societal inequalities.
The Public Sector Equality Duty and Bridges
This issue of algorithmic bias was the second successful ground of challenge in Bridges. Section 149 of the Equality Act 2010 imposes a Public Sector Equality Duty (PSED) on public authorities, including the police. It requires them, in the exercise of their functions, to have ‘due regard’ to the need to eliminate unlawful discrimination, advance equality of opportunity, and foster good relations between different groups.
The Court of Appeal found that SWP had failed to discharge this duty. The force had not taken sufficient steps to inquire into and satisfy itself that the risks of race and sex discrimination through algorithmic bias were being properly addressed. The evidence showed SWP was aware of the potential for bias but had not gathered the necessary information to assess the severity of this risk in the specific system it was deploying. As the Court stated, SWP "never sought to satisfy themselves, either directly or by way of independent verification, that the algorithm was not suffering from any such bias" (Bridges, [136]).
This failure is critical. The PSED is designed to ensure public bodies proactively consider equality rather than waiting for problems to arise. By deploying a novel and potentially discriminatory technology without a thorough prior assessment, the police failed in this duty. This aspect of the Bridges judgment confirms that, as a matter of UK public law, the way in which LFR was being used was incompatible with statutory equality protections. The judgment sends a clear message that a ‘tech-first, ask questions later’ approach is unlawful.
Combined with an Article 8 claim, this also raises the spectre of a violation of Article 14 ECHR, which prohibits discrimination in the enjoyment of Convention rights. If the interference with privacy under Article 8 falls more heavily on certain racial groups or genders due to algorithmic bias, and there is no objective and reasonable justification for this difference in treatment, it would constitute unlawful discrimination. The PSED failure in Bridges would serve as powerful evidence in supporting such a claim.
The Adequacy of the Legal Framework and the Path Forward
The Bridges decision did not impose a permanent ban on LFR but laid bare the profound inadequacy of the legal framework governing its use. The judgment effectively created a legal and political imperative for reform. However, in the years since, the government's response has been muted, and no primary legislation specifically addressing LFR has been forthcoming. Police forces have updated internal policies and conducted Data Protection Impact Assessments (DPIAs), but this does not cure the fundamental defect identified by the Court of Appeal: the absence of a clear, democratically-mandated legal basis passed by Parliament.
Reliance on the general provisions of the DPA 2018 and the UK GDPR is insufficient. These laws govern data processing generally but are not designed to regulate the specific, high-risk nature of state-operated biometric surveillance in public spaces. A DPIA is a procedural tool for risk assessment, not a substitute for a substantive legal basis for the interference itself (Oswald, 2021). The law must not only permit the processing but also set clear limits on its use.
The Ada Lovelace Institute has compellingly argued for an explicit legislative framework, suggesting that anything less fails to meet the ‘in accordance with the law’ standard (Ada Lovelace Institute, 2019). This framework would need to address, at a minimum:
- The specific and serious purposes for which LFR can be used.
- The criteria for placing an individual on a watchlist, subject to independent authorisation.
- Strict limits on where and when LFR can be deployed.
- Robust independent oversight and public transparency regarding its use and effectiveness.
- A mandatory requirement for independent testing and verification to ensure algorithms do not produce a discriminatory impact, in line with the PSED.
Without such a bespoke legislative scheme, the wide discretion that the Court of Appeal found unlawful in Bridges persists. The current situation, where police forces continue to develop and trial this technology under deficient general policies, represents a direct challenge to the rule of law.
Conclusion
In conclusion, live facial recognition, as it is currently being deployed in the United Kingdom, is incompatible with both Article 8 privacy rights and fundamental equality protections. The technology’s capability for mass, indiscriminate biometric scanning represents a profound interference with the right to privacy in public life, creating a chilling effect on civic freedoms. The judgment in R (Bridges) v Chief Constable of South Wales Police provides a definitive judicial assessment, confirming that the current legal framework fails the fundamental ‘in accordance with the law’ test. The patchwork of common law, data protection statutes, and internal policies is not specific, foreseeable, or robust enough to prevent arbitrary interferences with Article 8 rights.
Furthermore, the proven risk of algorithmic bias, resulting in discriminatory error rates against women and ethnic minorities, places the technology in direct conflict with the Public Sector Equality Duty. The failure of South Wales Police to adequately assess this risk before deployment, as found in Bridges, underscores a systemic failure to safeguard equality. This technological reality, when combined with a deficient legal framework, creates a system that is not only an affront to privacy but is also inherently discriminatory in its application. While it is theoretically possible that Parliament could design a new, specific legislative regime that renders LFR compatible with these rights, no such framework exists. Until it does, the deployment of this powerful surveillance tool operates in a legal and ethical vacuum, making it functionally and legally incompatible with the standards expected in a democratic society that respects privacy and equality.
References
Ada Lovelace Institute. (2019) The Biometric State. Available at: https://www.adalovelaceinstitute.org/report/biometric-state/
Big Brother Watch. (2018) Face Off: The lawless growth of facial recognition in UK policing. Available at: https://bigbrotherwatch.org.uk/wp-content/uploads/2018/05/Face-Off-final-2.pdf
Fussey, P. and Murray, D. (2019) Independent Report on the London Metropolitan Police Service’s Trial of Live Facial Recognition Technology. Human Rights, Big Data and Technology Project, University of Essex. Available at: https://www.essex.ac.uk/research-projects/the-human-rights-big-data-and-technology-project/reports-and-publications.
Information Commissioner’s Office (ICO). (2019) Investigation into the use of live facial recognition technology by law enforcement in public places. Available at: https://ico.org.uk/about-the-ico/news-and-events/news-and-blogs/2019/07/statement-on-the-high-court-s-judgment-on-the-use-of-live-facial-recognition-technology/
Malone v United Kingdom (1984) 7 EHRR 14.
National Institute of Standards and Technology (NIST). (2019) Face Recognition Vendor Test (FRVT) Part 3: Demographic Effects. NISTIR 8280. Available at: https://nvlpubs.nist.gov/nistpubs/ir/2019/NIST.IR.8280.pdf
Oswald, M. (2021) ‘Live Facial Recognition in Public Space: In Accordance with the Law?’, The Modern Law Review, 84(2), pp. 411-439.
R (Bridges) v Chief Constable of South Wales Police [2020] EWCA Civ 1058.
S and Marper v United Kingdom (2008) 48 EHRR 50.
Von Hannover v Germany (No. 2) (2012) 55 EHRR 15.
Woods, L. and Dowdney, L. (2019) ‘Live facial recognition: a step too far?’, UK Human Rights Blog. 13 May. Available at: https://ukhumanrightsblog.com/2019/05/13/live-facial-recognition-a-step-too-far-lorna-woods-and-laura-dowdney/
