The steady march of technological progress has finally shone its light into the dingy world of the mugshot – the commonplace name given to the photograph of a police suspect. The mugshot represents one of the earliest uses of technology in the identification of criminals. It has proved to be a particularly useful way of identifying recidivists – people who repeatedly reoffend.

Unhelpful inconsistencies in photography led to the implementation of standard anterior and lateral profile photographs taken at one-seventh scale – a convention that hasn’t changed in more than a century, even following the introduction of digital photography.

Dermatoglyphic fingerprinting and, ultimately, DNA profiling has offered more reliable forms of forensic human identification, but identifying faces remains critical in recognizing suspects. Police in England and Wales – using photo recognition technology – have now accumulated millions of images of people, many of whom may never have been charged with or convicted of any offense.

The case of S and Marper versus the UK in the European Court of Human Rights drew attention to the issue of retention of DNA samples and profiles of people who had never been convicted of an offense. It led to the UK systematically disposing of arrestee DNA and dermatoglyphic fingerprint evidence in routine cases not proceeding to a charge.

But what about the mugshots? These are no longer stored solely in an album on a dusty shelf in the corner of a local police station, but are digitized and searchable – and imminently there will be a capability for national remote database interrogation.

Others have drawn attention to the benefits, however, and security cameras are widely accepted by the public. Their value in a criminal investigation is unquestionable – over 3,000 suspects were recognized by eyewitnesses from footage examined following the London riots of 2011.

There are also dangers associated with the automatic computerized searching of CCTV video streams for people depicted in arrestee photographs. Should the mugshot databases be purged in much the same way that DNA and fingerprint ones are?

Public area computerized facial recognition has been just around the corner for decades – and still is. It’s worth noting that only one suspect in over 3,000 from the London riots was recognised by computer software and an attempt by “digilantes” to identify suspects by matching them to social media images held on the web failed.

The story of facial recognition software being used to identify a suspect at the 2017 Notting Hill Carnival was a similar debacle. That particular attempt apparently led to a number of false positive matches and people mistakenly being stopped by police.

Removing people who were arrested but never charged from police fingerprint and DNA databases exemplifies the uncertain consequences of dogged social action. The theoretical risk of a person who was never charged being associated with a crime scene they were never at has been reduced.

Probability dictates that detection will have been missed – including in serious offenses – simply because the database is smaller; although repeat offenders do predictably find their way onto the system.

Both the House of Commons science and technology select committee and the Biometrics Commissioner, Professor Paul Wiles, have expressed concerns about the delay in publication of the Home Office’s heavily delayed biometrics strategy. 

The government’s combined strategy needs to focus on the effectiveness of those biometrics that are reliable and on the contexts in which they work best – typically fingerprints, DNA, and irises. Eyewitness facial identification – especially of unfamiliar faces – is not reliable given that it underlies many serious miscarriages of justice. And facial recognition technology performs extremely poorly in public areas.

The pruning of the fingerprint and DNA databases at the hands of well-meaning civil liberty activists has led to two of the most reliable forms of forensic human identification being displaced by one of the least. It’s yet another perverse result for those interested in promoting reliable biometric and forensic identification to help the police do their job.