We have been asking the Government to clarify exactly what it means by the terms "fingerprint biometrics" with respect to the proposed compulsory National Biometric Identity Card / centralised Database scheme. Instead of publishing a full technical architecture of what exactly they intend to create, which would have no adverse impact on any commercial procurement process, they just keep dribbling out crumbs of information, apparently as an afterthought:
House of Commons Written Answers 28th June 2005 (pt 18)
"Mr. Austin Mitchell: To ask the Secretary of State for the Home Department how many digits he expects to be printed on identity cards; and what information he has collated on the number of digits required on passports and identity cards by other countries. [6164]28 Jun 2005 : Column 1452W
Andy Burnham: No fingerprints will be printed on the face of the identity card. Biometric information, including under current plans 10 fingerprints, will be held securely on the National Identity Register. Some of the fingerprint data could be held in an encrypted form on the card chip. All the Schengen states will be required to use biometrics in passports under Council Regulation 2252/2004. Fingerprint biometrics (rather than just fingerprints) will be introduced within three years of adoption. Non-Schengen states may choose to follow the requirements, although they would not be bound by the timetable. Other EU member states which issue identity cards are considering introducing biometrics to increase security and some, for example Greece, Italy, Portugal and Spain, already collect one of more fingerprints as part of their national identity card schemes. Among non-EU states, Hong Kong and the Philippines currently include fingerprint information on their identity card schemes."
Does
"Biometric information, including under current plans 10 fingerprints, will be held securely on the National Identity Register. Some of the fingerprint data could be held in an encrypted form on the card chip"
mean that the full fingerprint image scans are going to be stored on the National Identity Register, rather than any biometric secure hashes or template minutiae (an encoding of the ridges and intersections of the fingerprint pattern) i.e. the most useful data to both police investigators and to identity thieves will be stored centrally and have to be sent securely over the network infrastructure ?
There is no assurance that any of the fingerpint data will be held on the central National Identity Register in "encrypted form".
What does Andy Burnham mean about "some of the fingerprint data" with respect to the actual ID Card itself ? Does this means template minutiae of all ten digits or does it mean full digital scan images of, say, two fingers ?
We find it hard to believe that the Smart Cards being considerd for the ID Card scheme will have enough tamper resistant memory capacity to store the raw fingerprint image scans.
It is certain, that the ICAO standard Biometric passport chips (64 Kilobytes of memory in total) will not have enough tamper resistant memory space.
The current National Automated Fingerprint Identification System (NAFIS) for Criminal Records also uses 10 fingerprints and both palm prints (if you have them all) so you can see just how skewed the ID card scheme is towards treating us all as criminals, rather than as innocent citizens, and how far removed the system is from being an individual citiizen centric, purely identification card, and instead is really tool for mass surveillance by the state.
We fully support Dr. Brian Gladman's analysis of past abuses of the fingerprint system by the Police, and his fears for similar abuses of the ID card scheme.
[UPDATE: see comment below]
If four fingerprints are what the National Physical Laboratory / BT study suggested were needed for reasonably accurate identity verification purposes, and if even the US-VISIT system only uses two fingerprints for border control purposes, than surely the taking and indefinate storage of ten fingerprints of millions of innocent people must be considered to be disproportionate under the Data Protection Act ?
The NuLabour Home Secretary Charles Clarke seems to expect us to believe his statement during the Second Reading debate that
"I argue that the ID card system is in fact a bulwark against the surveillance or Big Brother society, and not a further contribution to it."
Is it terrorists, serious organised criminal, illegal immigrants, benefit fraudsters or "identity thieves" (the presumed statutory targets of the Identity Cards Bill) who are setting up multi-billion pound technological mass surveillance infrastructures to spy on everyone in the UK, or is it the NuLabour Government which is planning to do so ?
There is a good critique by Muriel Gray of the premise that the innocent have nothing to hide, see here:
http://www.guardian.co.uk/comment/story/0,3604,1519574,00.html
There are also agents of the state who will be put at risk because of the instance on the use of "biometric identifiers".
Undercover Policemen, James Bond secret agents, SAS special forces soldeirs etc. will all be issued with an ID card and have their Biometric identifiers registered at the age of 16, i.e. well before most of them will have been recruited into anuy of these sensitive posts. No doubt the Government can issue them with a ID card or Passport in a false name, but, if they have ever been overseas on holiday or on business to the G8 and EU countries that thewe are told are so keen to use biometrics for their passport and border controls, then these categories of peole will also be flagged up as "suspicious" and could have their "undercover roles" compromised.
If employers are meant to be able to "verify" that they are not employing "illegal workers" through the use of biometric checks against the National identity Register, then serious organised criminals will have access to fingerprint and iris scan and facial recognition information about undercover police or customs investigators, through their legitimate "front" companies, on the basis of "checking out" a potential employee.
The question only asked about fingerprints, not about palm prints, so it's not clear whether palm prints will also be taken. Obviously if the plan is to use palm prints, the minister should have said so in answering this question; but the way the government are letting information trickle out makes me unsure whether he would have.
Perhaps secret agents and so forth are meant to subvert biometric checks with gelatine patches on their fingers, special contact lenses, etc. None of the biometrics are known to work either for a scheme on this scale or when most tests are carried out by operators with limited time, training and motivation. Border guards could be properly trained and motivated, but will have very limited time per passenger. Most other operators will have neither the time, nor the training, nor the motivation.
Of course, people who try to subvert biometric tests will sometimes be caught. So agents of the state may indeed be placed at greater risk. Moderatly well organized criminals are likely to be inconvenienced perhaps a little, perhaps hardly at all. But the easiest people for Big Brother (which could mean agencies of the State or private corporations) to track will be ordinary, generally law-abiding people.
Here are some points on the issue of storing full fingerprint images, versus only storing templates extracted from full images (of minutiae or otherwise).
By discarding full images (on central storage) the Government would remove the opportunity to benefit from technological improvements that required a different template format or analysis. That is without re-registering the whole population. Thus it would not be prudent for the Government to discard full images.
The issue of on-card storage, as full images or templates, is somewhat different. For a UK-only system, template storage may well be adequate. This would limit Point-of-Use (PoU) checks to the matching approach(es) selected at card creation time. However, for international usage, eg within passports, current and (I believe) upcoming international standards call for a level of interoperability that uses fingerprint images as the common denominator. There are additional standards for a small number of template formats (and ISO/IEC are currently looking to standardise 3 such formats). Acceptance of non-image template formats for international use (especially in passports) is, in my opinion, less likely to be put into use, because of problems with interoperability between matching algorithms. However, adequate interoperability may become technically viable in due course.
Obviously, the other issue with on-card storage is template size. In my view, storage of just 2 fingerprint images (or perhaps only one) is likely to be the chosen way, for PoU checks. This is on grounds of "adequate" performance together with shorter transaction times and single finger capture equipment having lower cost. Use of greater numbers of fingerprint images is likely to be essential only for detection of multiple applications during enrolment, and perhaps for other special purposes (eg forensic), where access would be made to full 10-print sets stored centrally.
Best regards
The Third Principle of Data Protection, under the Data Protection Act 1998:
http://www.opsi.gov.uk/acts/acts1998/80029--l.htm#sch1ptI
"3. Personal data shall be adequate, relevant and not excessive in relation to the purpose or purposes for which they are processed."
If only one or two or four fingerprint biometrics are required for the purpose of a Passport, there is no justification for collecting 10 fingerprints when you apply for or renew a Passport, especially during the so called "voluntary" phase of the legislation, before Parliament has voted on "compulsion" to register on the National Identity Register.
Similarly if only a digital image and a thumbprint is required for a Passport, then there should be no compulsion to provide iris scans, for that purpose, which are not even going to be used on the embedded RFID chip.
Any new biometric image processing techniques requiring a different incompatible format of template stored on the physical ID card itself, are unlikely to be so staggeringly better than the current ones, that a "big bang approach" would be justified to upgrading them. Surely that can be done after a certain date with new or renewed ID cards or passports ?
If a template format is found to be easy to forge, or the digital signature master keys are compromised (e.g. the correct SHA-1 secure hash collision is found, with or without inside knowledge) then the Government must factor in the risk of the costs of having to re-register millions of people in a hurry.
More likely they will try to ignore any such systemic failures and hope that they will not be too widely exploited, until re-registration with a new algorithm, or key etc.
By email from Dr. Brian Gladman:
"Hi
Thank you for your kind reference to my paper concerning the potential misuse of fingerprint data held in the proposed UK National Identity Register
In order not to confuse your readers, I should perhaps point out that the way in which you comment on this:
"We fully support Dr. Brian Gladman's analysis of past abuses of the fingerprint system by the Police, and his fears for similar abuses of the ID card scheme."
implies that I covered past fingerprint abuses when in fact my comments on the past were actually about DNA records, not fingerprints.
I am not trying to be critical but just pointing out that this might confuse people if they look at my paper expecting to find coverage of past abuses of fingerprint data.
Of course its just as bad that the Police are allowed to keep the fingerprints of those who are innocent - but I am not aware of the history of this - was it ever illegal to do this?
best regards,
Brian"
wtwu commented above, at July 3, 2005 11:18 PM (that PM is puzzling): "[number of biometric modalities/instances must be] adequate, relevant and not excessive in relation to the purpose."
I do agree that this (and other) aspects of the data protection act have a good basis.
However, extra biometric samples (same or different modality or extra instances, eg 2nd iris scan and extra fingerprints) do actually improve performance (false accept/reject trade-off); see eg http://www.camalg.co.uk/s05011_tr0/s05011_tr0.pdf
This is very important for Detection of Multiple Applications (DMA). Given that DMA is a legitimate aim of the proposed National Identity Scheme, it strikes me as therefore almost impossible to argue logically that the number of biometric samples taken can ever be excessive or disproportionate in an absolute way.
One could argue against taking many biometric samples on grounds of enrolment time/effort, and on the cost of the number of different biometric equipments required at each enrolment station.
There is also a case for limiting the number of templates placed on ID cards to be less than the number available on-line. In particular, in addition to costs of card storage and point of use biometric devices, this is to prevent disclosure of all templates to the risk of artefact attack (eg gummy fingerprint overlay), so holding some in reserve for on-line checks.
Perhaps, even more than the above points, wtwu's underlying concern is that it is wrong to take 10-print sets of fingerprints of the whole population, for forensic use at the whim of any policeman. That is an entirely different matter, where I am in agreement. My view on this is that the police should have access to the NIR, for general searches of this sort, only with a warrant from a high court judge. It is also likely to be desirable for a more senior police officer (eg chief inspector or superintendent) to be able to authorise such checks of the fingerprints of an unidentified and uncooperative person (thought to be UK resident) who has been arrested on a serious charge.
I also recognise that a legitimate case can be made for excluding fingerprints entirely (or above any specified number of "flat prints") to prevent government/police misuse of fingerprints in the NIR for forensic purposes. Currently that is not a view that I hold, but I do see it as a view that is not illogical.
Best regards
I understand that it might seem to be a good idea to keep some literaly "check digit" fingerprints centrally which are not stored on the ID smartcard. However, in real life, how often is a situation likely to occur when this extra complexity and functionailty is likely to be used ?
What situation would ever trigger a request for, the check of say, the left hand little finger print, if for example, both thumb prints are what are usually checked ?
If the two thumb prints check out ok against what is stored on the ID card, then any extra online check for the little finger print is only going to cause confusion, delays and queues.
If one or both thumb prints "fail" against the
stored biometrics on the Card, then why would you not check them first against the central database ? If the presented thumbprints fail against the centarl database, then surely a check against the little finger print is superfluous ?
I do not understand why the Home Office does not take the same approach as with the Regulation of Investigatory Powers Act, regarding the authorisation and oversight of electronic phone and email intercepts.
The way the Identity Cards Bill is written implies that the even the most junior member of staff in the police or intelligence services can be authorised to have full read access to the National Identity Register, including the detailed audit trail, assuming that the named roles e.g. Director General, Chief Constable etc. of the heads of these organisations are allowed to delegate any such tasks to their subordinates.
Given that biometrics do not offer perfect performance, from time to time an enrolled user will be rejected on the basis of the first biometric check (say 1 or 2 thumbprints).
In this case, a secondary check should be provided (by biometrics or otherwise) and in a way that does not blame or discriminate against the claimant. Checking additional fingers (perhaps on-line) is one form of secondary check.
Checking an additional finger provides additional (rather than replacement) information. Thus the overall result, using the extra fingers, should be more accurate. It would, however, take longer - so is best avoided, except where necessary.
There is also the possibility that the first digit(s) checked are damaged or changed from enrolment, temporarily or permanently.
Checking prints against on-card templates and then against centrally held templates is not useful, unless different templates are stored, or different matching algorithms are used. Although this is possible, it is not particularly beneficial to set up the system that way.
Unfortunately, I have not understood wtwu's point on the RIP Act. My reading of the draft ID bill is the same as wtwu's, that access to the NIR is not adequately restricted concerning authorisation of access by the police and intelligence services.
Best regards
Under RIPA the organisations who have access to your data without your permission are listed, and the authority to ask for phone or other electronic intercept warrants, or for communnications traffic data is laid down according to diffrent, broadly similar ranks of police or intelligence officer etc. e.g. deputy chief constable or above or superindent (for traffic data).
There are well established procedures for a SIngle Point of Contact within each police force or intelligence agency who filter out any excessive or repetitive or illegal request, so as not to waste either the Home Secretary's time or that of the busy commercial organisations who have to comply with these requests (in secret).
Under the Identity Cards Bill, these police and intelligence agencies are only mentioned by reference to the Director General of MI5 the Security Service, the Chief Constable of a Police Force etc. and there are no statutory minimum ranks or roles laid down regarding who can request your private data without your permission.
So, instead of requiring permission from a Judge or even from the so called independent Identity Scheme Commissioner to access a particular person's private National identity Register data without their permissin, the Identity Cards Bill is even more lax in the number of people that could legally request such data. Obviously there may be internal codes of practice, but these are not legally enforcable, and, in any case, the Identity Cards Bill provides a "good faith"/"I thought I was properly authorised" defence against being prosecuted for any illegal discloures.
There do not seem to be any safeguards against speculative trawling or data mining of the entire National Identity Register, or even to prevent cloning it onto another system outside of even the limited regulatory protection offered in the Identity Cards Bill, by one of the authorised police, intelligence, or even , in the future, licensed credit reference agencies etc.
How exactly will it be possible to resist feeding the NIR biometric data into the European Union's Schengen Information System II ?
http://www.sirene.gov.uk
The Identity Scheme Commissioner looks to have even fewer powers than the various RIPA or Data Protection Commissioners
http://www.spy.org.uk/cgi-bin/commissioners.pl
who are somehow meant to protect our private data from being abused.
Nigel said it more succinctly:
"access to the NIR is not adequately restricted concerning authorisation of access by the police and intelligence services."
sir
I am final year B.TECH student.now i am doing on my final year project(fingerprint door opening system).i donot hava enough knowledge so please send fingerprint compare information.