Article

Facial recognition in policing: part one

How the UK and US are taking divergent paths

At Altia, we work with law enforcement agencies around the world, giving us first-hand insight into how attitudes towards technology in policing differ by region. Recent developments highlight this divergence clearly. In the UK, the Government’s Policing Reform White Paper identifies live facial recognition (LFR) as a modern capability more forces should adopt. In the US, meanwhile, federal agencies are expanding their use of biometric tools amid ongoing scrutiny and debate. Together, these developments illustrate two distinct approaches to the same technology.

A shared technology, different philosophies

Facial recognition technology (FRT) has moved rapidly from research into frontline policing. It enables the comparison of facial features from images or video against datasets to help identify individuals, supporting investigations, locating suspects, and preventing harm at public events.

While the technology itself is global, attitudes towards its use are not. Both the UK and the US face similar challenges: balancing public safety with civil liberties and maintaining public trust. However, each has adopted a different philosophy for deploying facial recognition, shaped by governance structures, legal traditions, and public expectations.

These differences matter. For police leaders, policymakers, and technology providers, how facial recognition is governed influences not only operational effectiveness but also long-term legitimacy and confidence in policing.

The UK approach: coordinated deployment with guardrails

In the UK, facial recognition is increasingly framed as a legitimate policing tool, provided it is used responsibly and transparently. National policy emphasises LFR as a targeted capability rather than mass surveillance. Guidance and funding from the Home Office support a coordinated approach across forces, underpinned by technical standards, oversight mechanisms, and clear governance frameworks.

Court judgments and regulatory guidance have reinforced requirements around necessity, proportionality, and data protection. Public communication is central: forces are expected to publish impact assessments, explain watchlist construction, and demonstrate that officers retain decision-making authority. Facial recognition is positioned as an investigative aid, not a replacement for human judgement.

Deployments are typically framed around preventing serious crime and protecting public safety, such as locating individuals wanted for violent offences or safeguarding vulnerable people. This aligns with the UK principle of policing by consent, where legitimacy is derived from public confidence that powers are exercised responsibly.

The US approach: fragmented adoption and rights-first scrutiny

The US has no single national strategy for facial recognition in policing. Governance is fragmented across federal, state, and local levels, reflecting decentralised law enforcement structures. Federal guidance often emerges through Department of Justice policies, oversight audits, and civil rights analysis.

Rather than encouraging widespread adoption, US debate centres on restraint. Some cities and states have banned police use of facial recognition entirely, while others allow limited deployment under strict conditions. Where used, it is typically accompanied by privacy policies, audits, and restrictions on specific use cases.

This patchwork reflects deep-rooted concerns about civil liberties, constitutional protections, and the potential for technology to exacerbate existing inequalities. As a result, facial recognition is treated less as an operational necessity and more as a high-risk capability requiring continual justification and reassessment.

Ethics, trust, and common ground

Despite these differences, ethical concerns are strikingly similar on both sides of the Atlantic. Issues around algorithmic bias, privacy intrusion, and over-reliance on automated systems remain central. At the same time, the potential benefits are clear: when used responsibly, facial recognition can accelerate investigations, prioritise resources, and provide analytical scale that would otherwise be impossible.

Trust depends on how the technology is embedded in policing practice. Transparency, public engagement, and rigorous human oversight are essential. Facial recognition must remain a supportive tool, with accountability for outcomes clearly maintained.

Altia’s perspective on AI

At Altia, our approach to AI is guided by a simple principle: technology should enhance human judgement, never replace it. In policing environments, tools only create value when they are trustworthy, proportionate, and designed with a clear understanding of their social and operational impact.

We prioritise disciplined innovation over speed, focusing on real-world behaviour, potential failure modes, and how technology reshapes decision-making. Our systems are designed so investigators remain in control, with automated analysis supporting, not supplanting, human authority and accountability.

We emphasise fairness, robustness, transparent governance, and strong privacy boundaries. Our tools are built to be explainable, contestable, and proportionate, reflecting the expectations placed on modern policing.

Responsibility first

The UK and US may be taking different paths, but responsibility is their shared foundation. Facial recognition is not just a technical capability; it reflects broader questions about power, trust, and accountability in the digital age. For policing organisations and technology providers alike, success is defined not by how quickly systems are deployed, but by how carefully they are governed and legitimised.

That is how Altia approaches AI; not by chasing hype, but by ensuring technology earns its place and genuinely serves those who rely on it.

Core differences at a glance: