Thursday 24 May 2012

Police forces all over the UK are introducing mobile fingerprint equipment. Result? Approximately 20% of the criminals who would otherwise have been taken down to the station will now be asked politely to go on their way

The Guardian tell us today that the Metropolitan Police have bought themselves some new equipment – mobile fingerprint readers. They are the 25th UK force to do so.

It all seems very sensible:
One of the aims of the technology is to cut the number of trips police make to the police station, so that officers can spend more time on the frontline.

Mark Rowley, assistant commissioner at the Met, said: "Evidence has shown that a full identification arrest can tie-up both the subject and the police officer for several hours. Even a traditional identity check conducted on the street can take an extended period of time to complete.

"It is effective particularly in revealing serious and violent offenders who will do everything they can to prevent the police from knowing their true identities."
It isn't.

Because the failure rate on this technology is about 20%*.

In the case of people whose fingerprints are on file, about 20%* of the time that fact will not be discovered using this technology.

To be slightly more technical:
  • When the Home Office tested flat print fingerprinting in the UKPS biometrics enrolment trial back in 2004 they found that the false non-match rate was 19 or 20 percent*. Nick Herbert, say, would be told by the system that he was not Nick Herbert. His prints didn't match anything on the database. A non-match. A false non-match, as it happens, as Nick Herbert had just registered his prints on the database five minutes before.
  • That's how the trial was conducted. 10,000 of us registered our faces and our fingerprints and our irisprints and five minutes later we tested to see if we could have our identity verified using one or other of those biometrics. Flat print fingerprints failed 19-20 percent of the time. Face recognition failed between 31 and 52 percent of the time which is why the smart gates at our airports and every other instance of automated face recognition are guaranteed to be a waste of time and money.
Nick Herbert is the policing minister. He and the Home Office have not published the results of any flat print fingerprint trials since then. At least one trial has been performed:
The NPIA signed a contract with 3M Cogent in 2010 for mobile fingerprint identification devices. The deal followed field trials involving 28 police forces using Lantern devices to test how mobile fingerprinting performed in an operational environment.
But the Home Office refuse to publish the results. So as far as we know – we, the public – the false non-match rate remains approximately 20%.

Which means that out of all the wanted criminals who are stopped and whose prints are on IDENT1 – the national fingerprint database – 20% will falsely not match and be allowed to go on their way.

There's a question for Mr Herbert forming at the back of your mind, isn't there – do you know what you're doing?

If he wants to prove that the answer is yes, and that he's not undermining the fight against crime and wasting our money at the same time, then Mr Herbert must publish the Lantern trial results. Nothing else will be convincing.

The chances of the Home Office publishing those results? See yesterday's hell-freezes-over press release.

It's not just DMossEsq:

* Please see UK Passport Service Biometrics Enrolment Trial Report May 2005, Management Summary, Key Findings, para., p.10:
Fingerprint verification success

• The majority of participants achieved successful verification on fingerprint, with rates of 81% for Quota participants and 80% for Disabled participants. One of the factors influencing failure was that the single fingerprint device used for verification occasionally did not record sufficient detail from the fingers.

• Younger participants had a higher fingerprint verification success rate than older participants.

Updated 20.2.18

It's nearly four years since the blog post above was published. Not a single success story for mobile fingerprinting has been told. Four years. Zero results.

Unabashed, Yorkshire cops have begun using on-the-spot fingerprint scanners.

The difference this time is that the policemen on the street will be able to interrogate not just IDENT1 – the national criminal fingerprint database – but also IABS, the Immigration and Asylum Biometrics System: "The scanners link up to an app on cops' smartphones – which is already available to all 5,500 frontline officers – and run the prints against the UK's criminal fingerprint and biometrics database (IDENT1) and the Immigration and Asylum Biometrics System (IABS)".

What is unchanged is the 20% figure. It remains the case that 20% of those stopped by the police will be falsely not matched. Thanks to this flaky biometrics technology, wanted criminals will be asked to move along when actually they should be detained.

No comments:

Post a Comment