The Musings of Harry

According to a recent YouGov poll (hat tip: Ideal Government), 75% of people think that lots of CCTV is a good idea; 50% would like to see a universal police DNA database; 55% would like centralised NHS records and 45% of people support mandatory ID cards… yet only 35% of those polled thought that it was a good idea for the government to collect more personal data.

As far as I’m concerned, this is clear evidence that people simply do not understand the implications of these schemes. It would be interesting to see two versions of this poll conducted: one where they ask the personal data question first, and another where they ask it last. I would bet good money that the first poll would show much less support for these schemes than the second — a small reminder can work wonders.

Perhaps a few more HMRC-esque data catastophies will place more firmly in people’s minds the risks attached to the aggregation of massive quantities of private data…

This story is a perfect illustration of the stupidity of the Police’s current DNA retention policy.

The story is pretty simple: someone mistook a man’s MP3 player for a pistol (I’ll leave aside that that’s probably really stupid…) and called the police. Armed police then arrested the man, placed him in custody and took his DNA and fingerprints. As soon as the mistake was uncovered, the man was released and the police apologised.

So far, so good. These things happen; it’s unfortunate but unavoidable. His DNA and fingerprints, however, will be retained, along with a file noting that he was arrested on suspicion of a firearms offence. This is not good at all.

It is, in fact, really bad. It’s manifestly illiberal, but also bad on entirely practical grounds: it clutters their database with useless records. A clean database full of the DNA of criminals is clearly a useful resource; a messy database full of the DNA of everyone you happen to meet is much less useful. It increases the chances of obtaining false positives and creates additional, spurious leads that must then be followed up by officers. This is all the more true when techniques with errors rates known to be high, like low copy number testing, are used.

Civil liberties concerns aside, there might an argument in favour of adding the DNA of people you strongly suspect but cannot prove to have committed an offence — but adding the DNA of those who are clearly innocent is just eye-rollingly stupid.

Ever since the announcement by Alistair Darling in the Commons that the Government had lost 25 million people’s personal data, ministers have been spouting the fantasy that biometrics will prevent the same thing happening with the National Identity Register, which will underpin the ID cards scheme.

It is nice to see that someone has finally called their bluff. Academic ninjas Ross Anderson and colleagues have published an open letter to the Government explaining, in excellent language that even a minister could understand, why biometrics won’t make a jot of difference. This is splendid news, although I suspect it’ll be ignored along with the rest of the advice that academics have offered the Government over the last few years. Nevertheless, somebody had to do something to counter the stream of fantastical nonsense which has been flowing, unstemmed, from the mouths of Government ministers over the last week or so:

“The key thing about identity cards is, of course, that they will mean that information is protected by personal biometric information. The problem at present is that, because we do not have that protection, information is much more vulnerable than it should be.” — Alistair Darling, 20/11/07

This is a complete non sequitur. It makes no sense whatsoever to anyone with even the most frail grasp of the technology at work. The ignorance this demonstrates is appalling.

“What we must ensure is that identity fraud is avoided, and the way to avoid identity fraud is to say that for passport information we will have the biometric support that is necessary, so that people can feel confident that their identity is protected.” — Gordon Brown, 21/11/2007

Is it now? Is it actually safe given that in April — 7 months ago — academics in the states managed to reconstruct a fingerprint capable of fooling a scanner from the data which is sent from a passport to a reader over the air when its RFID chip is scanned? The very same data which, despite being encrypted, has such a weak key that it can be cracked in a few minutes by anyone with a desktop PC?

“There is of course an important protection in an identity card system, through the use of biometrics. Biometrics will link a person securely and reliably to his or her unique identity. It will therefore become much more difficult for people to misuse other people’s identity, even if full details of their biographical information are already known. The current plan for the national identity register is for biometric information to be held separately from biographical information, thereby safeguarding against the sort of eventuality that the right hon. Gentleman described.” — Jacqui Smith, 21/11/2007

In response to this pleasant-sounding fantasy, David Davis replied that he did not look forward to the day when somebody asks for this data and is sent it. Indeed. How, also, does Ms Smith propose that biometrics be used to secure information when a fingerprint, suitable for fooling a scanner, can be lifted from a tumbler or a CD case and attached, almost undetectably, to the tip of one’s finger?

These people are living in a complete fantasy, and a dangerous one at that. I don’t think the HMRC scandal is the nail in the coffin for the ID cards scheme, but it’s certainly one of them. Nor do I think that, were the ID cards scheme to die a quiet death, the problem would be solved.

This government has big plans for centralisation of data, and big plans for lots of neat technology, but they lack the most important ingredient: a culture of privacy protection and respect for people’s personal information. The absence of this trait is stark and worrying. What does it say about a government that hires expert consultants, for hundreds of thousands of pounds, to produce policies on data handling that fill inch-thick books with obvious advice, only to leave them languishing in a drawer while giving junior office lackies access to half the population’s personal data without any effective supervision or oversight?

Is it really prudent for this Government to create more and more and more and more databases when these incidents are happening over and over and over and over again? What planet are these people on? When will they be jolted out of this utopian technological fantasy?

If this hasn’t done it, what will?

Here, most unfortunately, we go again.

I recently posted about the partisan brain, and about how Neil Harding rather exemplifies it. It seems he was all too eager to offer up another example.

Mr Harding, it would seem, wishes to go even further than some police offers, who recently suggested that we should collect DNA for anyone who is arrested. He thinks that we should collect DNA from everyone (presumably at birth — who knows?) and keep it on file, just in case you commit a crime.

Harding says that the DNA database helps police catch the “correct perpetrators of many crimes”. He says that the success of the database is “unquestionable”. His use of words is interesting, considering that people are, in fact, questioning it: just last week, in Liberty’s response to the Home Office consultation on PACE, they asserted that the National DNA Database “does not seem to have a significant impact upon crime detection”. In fact, although the number of samples on the database has increased steadily since it was introduced…

…the rate of crime detected using the database has stayed at about 0.35% of all recorded crime (see Liberty response). If the database were so useful, one would expect that it would facilitate the detection of more crimes as it grew larger. That does not seem to be the case. It is, therefore, of questionable value. I’m sure it has a place, but collecting the DNA of everyone arrested, let alone everyone alive, is not it.

Another terribly important point to make is that all systems have a rate of failure. Although (bar twins) an individual’s DNA is unique, we are not necessarily able to collect, store or compare samples with sufficient precision to make a DNA profile unique to an individual. Just as with fingerprints, DNA profiling techniques do experience false positives and false negatives, and just as with fingerprints, these rates are more than one might expect. Here is an excellent page detailing many such experiences.

I’ve written about these problems before, so I’ll leave the subject there, except to say that a good way to deal with these problems is to keep the database as small as possible. By keeping the database small, one minimises the number of failures, making them more manageable. This, quite neatly, makes it eminently sensible to restrict people placed onto the database to those convicted of a crime: perhaps even only those convicted of a serious crime. The usefulness of the system is maintained, as is the privacy of innocent people: something which, quite clearly, Harding grossly undervalues.

PS: If the database is so fantastically useful, and if a person being arrested is a useful indicator of criminality, why didn’t Levy, Evans and Turner have their DNA sampled?

The Home Office’s review of the Police and Criminal Evidence Act was published today.

Having read it, I have to say that it is reasonably balanced. A wide range of views are expressed. One wonders what will happen next. Privacy International expressed the rather scathing view that the proposals being considered were vague and undefined, and that the Home Office was merely going through the motions in conducting a public consultation at all. One hopes that they are wrong. Here, in any event, are the things that caught my eye.

Powers of Arrest
Two points of note here. First is the eminently sensible suggestion that the citizen’s power of arrest should be extended to apply to any offence, not just to an indictable offence. If a citizen reasonably suspects that a crime is being committed, it is proper that they should be able to restrain the suspected offender until the police arrive. To expect a citizen to have sufficient knowledge of the law to know whether or not the offence is indictable or not is unrealistic.

Second is the depressingly predictable suggestion that PCSOs be given the power of arrest, the argument being that since they have the power to detain people for 30 minutes until real police arrive, they have de facto powers of arrest anyway. One must ask: if PCSOs are to be given powers of arrest, what differentiates them from real police? They must, if given powers of arrest, be “police on the cheap” as this blogger has suspected all along. The remit of PCSOs is different from that of real police (and, frankly, questionable) and the distinction is important. They are receive less training and less pay. They are not police officers, and it is not right to expect them to perform the same duties.

Collection of Biometric Data
Here we go again with crackpot suggestions from senior police officers. One Inspector Thomas Huntley responded to the review with the suggestion that the power to collect DNA should be extended to allow the collection of samples from people arrested for non-recordable offences: speeding, littering etc. In support of this suggestion, he offers the following, ridiculously bone-headed, argument:

“… failure to do so could be seen as giving the impression that an individual who commits a non-recordable offence could not be a repeat offender and therefore only those arrested for recordable offences are likely to have offended before.”

I reject his premise, utterly. The mere fact that a police officer chooses to arrest someone does not imply guilt; it does not imply that they are an offender at all, let alone a repeat offender. Retaining DNA samples only for those convicted of an offence should be the standard that we set for inclusion in the National DNA Database. ACPO may not agree with that, but even they think that sampling people for littering is a step too far:

But despite some support for the police’s request to massively expand the database, the Association of Chief Police Officers (ACPO) warned that granting such powers could contribute to the increasing “criminalisation of generally law-abiding public”.

On the same topic, Liberty threw down the gauntlet with a challenge to the assumption that the NDNAD is even useful: they say that statistically, it has had little impact upon the detection of crimes. Another encouraging suggestion came from the Human Genetics Commission, which expressed concern that DNA samples are held for innocent people, and suggested that those convicted of minor crimes should have their samples destroyed if they do not reoffend for a set period. Good stuff.

Search Warrants
Finally, the review reports that several respondents suggested that the power to issue search warrants should be removed from magistrates and conferred upon police superintendents! The suggestion is that the issuing of these warrants would then be reported to the courts after the fact. In an astonishing display of doublethink, one respondent from the CPS even claimed that such a move would increase accountability and transparency… This asinine suggestion is further evidence of a worrying trend: the reduction of the role of the judiciary in overseeing the actions of the police. Search warrants need to be approved before they are executed, not after it is too late to take action if a warrant is unreasonable.

When the time comes for the Home Office to announce what changes will be made to PACE — if any — it will be very interesting to see which of these suggestions are taken on board.