You will, perhaps, by now be getting a sense of the challenge of setting up an identification system on a national scale. However, for many routine purposes, establishing who a person is from an entire population of possibilities is not what is required. Instead what is required is confirmation that the person is who they claim to be. This is verification. An example of verification happens when you collect a parcel from a depot. You are sometimes asked to show your driving licence, passport or other suitable document. This document verifies your identity.
Verification, like identification, involves a comparison of data, at least when done properly. If I sign for the parcel when I collect it, my signature should be checked against the one in the verification document. Alternatively, if the verification document has a photograph of me, the person issuing the parcel should check me against the photograph.
What distinguishes verification from identification is the number of checks made. In identification, data I supply (or which is taken from me) is checked against an entire database of templates. With verification, my data is checked only against data in the verification document. To use the terms you have already met, identification involves one-to-many checking, whereas verification involves one-to-one checking.
Because verification involves just a single comparison, it offers some advantages over identification. One advantage is that a biometric template can be stored in the verification document itself, in a machine-readable chip. The Schiphol airport system shown in Figure 5 uses this method. Travellers who enrol in the system receive a membership card containing a memory chip. The chip holds the member's biometric template. When passing through the airport, the traveller puts the card into a reader and stands in front of an iris scanner. If the scan matches the template on the card, the traveller passes through immigration control without further formalities.
Another advantage of verification over identification relates to false positive identification. In identification, biometric data has to be checked against every template in the database. The large size of a national database can make false identification quite likely. However, in verification there is only a single check – against the template in the verification document. In fact, in verification we are much more likely to be concerned with false non-matches than with false matches. For example, an enrolled traveller at Schiphol airport would be highly inconvenienced if his or her iris scan did not match the template on the chip. This is a false non-match.
Naturally verification is only as reliable as the document used for verification. As the extracts from Dettmer's article have shown, enrolling users in the proposed UK national identity card scheme requires that the applicant be checked against an entire database of previous applicants before the card can be issued. In other words, enrolment depends on an identification system. Only when an applicant has been successfully checked against the entire database is the verification document issued. This feature distinguishes the UK identity card system from those of some other countries.
Click on 'view document' to read the final extract from Dettmer's article (with thanks to R. Dettmer).
Activity 23 (self-assessment)
Now read the final extract from Dettmer (2004), and answer the following question.
Dettmer foresees problems in setting up a national identity card scheme. Does he see these problems as relating to identification or to verification? Illustrate your answer with one or two short quotations. (There is no need to supply references for your quotations.)
The problems Dettmer mentions nearly all relate to setting up a national database for an identification system. For instance, he mentions uncertainty over 'how long it will take to enrol users in a real system' and 'what percentage of the population will be unable to travel to their local biometric registration centre'. He says very little about the use of the card for verification.
Dettmer's anxieties, as revealed in the article, relate to the practicalities of the system. Will there be so many identification errors as to make it unworkable and undermine the public's trust? Can people be enrolled fast enough? And so on. However, for many other critics, anxiety is based much more on the principles than on the practicalities. I shall look at some of these concerns now.