Do biometrics hold the key to prison release?

By Professor Fraser Sampson, former UK Biometrics & Surveillance Camera Commissioner

In the criminal justice setting there are two questions in that biometrics can answer better than any anything else:

1.     Who’s this?
2.     Whose is this?

The first of these is the original use of anthropometrics to identify repeat offenders reliably within the criminal justice system as pioneered by French police official Alphonse Bertillon in the late 1880s. The second arises in the use of biometrics in forensic science as revolutionised by fingerprint and DNA technology, for example, in crime scene management where something found at a scene (bodily fluids) can be matched with someone whose identity is already known.

The reason they are so good is because biometrics is a comparative science. Biometric data has little absolute value in the justice system, but where you have reliable comparators, it is exceptionally powerful. The UK’s DNA database produced 821,794 matches to unsolved crimes between April 2001 and March 2024 (an average of 35,730 per year). In answering both questions, DNA matching has a very high level of accuracy, with the greatest number of shortcomings in recent years coming consistently from human error.

Both criminal justice questions rely on comparison for their answer, looking for predetermined points of congruity. Both depend on precision of measurement and availability of comparators. Finding a way to answer the ‘who’s this?’ question, reliably, quickly and conveniently has preoccupied the biometric ID community for the past decade or more. Rapid advances in financial and professional services, where decisions based on identity demand a high level of confidence that you’ve got the right person, are behind the many know your customer (KYC) and anti-money laundering protocols (AML) now in common use.

From critical infrastructure protection to international aid distribution there are many scenarios where it is imperative to know that the person in front of you is who they are claiming to be or who you believe they might be. The latest biometric capability offering accuracy and precision in identification is live facial recognition (LFR) and the performance report just published by the Metropolitan Police shows how effective it has been, mainly in catching people who were unlawfully at large or breaching conditions of a court order.  

Against this backdrop, the recent crisis of prison releases in the UK looks like a problem with a tailor-made biometric solution. Before going on, I recognise that one thing the criminal justice system doesn’t need is more lazy solutionising. Of course, prison staff could use facial recognition technology (FRT) to confirm the ID of individuals and corroborate authorisation before treating them as they are proposing to do (release, remand, transfer). But turning the obvious into reality is the holy grail of policy and a biometric fix across the criminal justice system would take a few (big) things.

The beleaguered minister said of the latest UK release error that he wasn’t “equipped with all the detail” before answering to parliament – and that’s the point. No one’s equipped with all the detail.

I don’t underestimate the scale of the problem but, should they decide to make a case for an early LFR solution, the prisons sector has some clear advantages over policing.

Despite having demonstrated the benefits of FRT after a shaky start policing still faces some significant challenges. The intrusiveness of pointing the technology at the general public is one, along with concern about how ‘watch lists’ and databases are compiled, and disparities between environmental conditions under which algorithms are tested for accuracy and bias and the reality of the operational settings in which the police use them. All of these are fair points for policing to answer, and none apply in a custodial setting.

Unlike streets, shops, stations and other public spaces, prisons are designed to be places of enforced presence. With a known, finite population, powers of compulsion, spatial control and biometric enrolment capability where matching is already taking place, prisons and custody suites represent a wholly different LFR use case from public space surveillance. While there’s an element of monitoring and observation, in deciding whether someone can be released from lawful detention you’re not looking for people – you’re looking for confirmation. What that process appears to need is what biometrics does best.

The Metropolitan Police performance report illustrates how the biometric technology has been used to give police teams a fighting chance where their numbers are going down and those of the other side are going up. Even the self-defeating policy of advertising its use in advance so people can opt out of being captured by LFR cameras would not be applicable in prisons, nor would concerns around mission creep.  

We talk of a criminal justice ‘system’ but it’s really a cumbersome arrangement of disconnected legacy apparatus. A systemic solution would require all the data from courts, probation service, police and prisons to be synthesised to help frontline staff in critical decision-making, maximising effectiveness and minimising error.

An integrated system across the criminal justice sector will need serious groundwork, converting records into a comprehensive accessible database. It will need compatible infrastructure across the estate to allow for basic biometric verification so that decision makers are equipped with all the detail. But it could work.

ID verification is standard practice for many functions in our lives. We expect biometric controls at international borders or for sensitive area access because the banks and airports invested in the technology as part of an integrated security and assurance plan. Without such a plan across the criminal justice sector, we should also clarify our expectations of the status quo: we won’t get an enterprise application standard for ID verification with a pencil-and-paper security system even when it’s working normally. Where it’s compounded by “mounting pressures on staff and outdated administrative systems”, failure is all but guaranteed.

For the past 130 years Bertillon’s biometrics have proved very effective at putting the right people in prison; perhaps it’s time we thought about using them to let the right people out. 

About the author

Fraser Sampson, former UK Biometrics & Surveillance Camera Commissioner, is Professor of Governance and National Security at CENTRIC (Centre for Excellence in Terrorism, Resilience, Intelligence & Organised Crime Research) and a non-executive director at Facewatch.

Article Topics

biometric identification  |  biometrics  |  criminal ID  |  forensics  |  Fraser Sampson  |  law enforcement  |  prisons