Identification is a part of living and is something that is common to all living creatures. It is way to distinguish one being from another and without it; the world would be a mess, as no one would ever know who was doing what. Human beings however, have taken this to a whole new level, using it to classify and categorise people, as well as to limit the availability of information and resources. To this end, numerous methods of identification have been devised, many of which use fabricated identifiers to ensure uniformity and distinctiveness. This article will endeavour to explore the differences and similarities between two methods of identification. The first of these is Voice recognition, a method that uses a naturally unique attribute to ascertain the identity of a person. The second is the National Insurance Number, a unique, uniform code that all UK citizens are issued.
These methods of identification are very different in terms of both scope and tangibility. Voice recognition is used by everyone all the time, although we don’t really think about it, for example when a person rings a friend or acquaintance, both will use voice recognition to identify one another. However, it can also be used in a more mechanical way, an example of which would be the intelligence networks of various governments. The NSA uses voice recognition as part of the system that takes the information from all telephone calls and processes them (Poole, 2000, para. 2). Technology is used to systematically analyse the phone calls and identify certain people using voice recognition, enabling them to find them or to verify that the call was made by that person. So the range of applications for voice recognition is huge and as such is used in many different industries.
National Insurance numbers on the other hand are issued by the government for specific applications, such as identifying who is being paid what and if you don’t have one, by what authority and conditions you are in this country. In many ways, this method of identification is more clear-cut and easier to authenticate, i.e. has a higher confidence level, as it is a number unique to an individual with no overlap. Whereas voice recognition can be hazy, if 2 brothers sound very alike for example, it only takes little more than a little background noise to confuse a system or a person and in the case of computerised systems, 2 similar voices may be considered exactly the same or with the acceptable tolerances of the system. However, despite these problems with voice recognition, a number is easier duplicate than a voice, i.e. writing it down. These problems highlight the fact that no one identification method is truly the best overall, but instead more suited to certain functions.
The quality or reliability of authentication methods for National Insurance numbers depends on the nature of its application. For example, for the purposes of getting a job it is useless on its own, as anyone could make a card with a number on it and will therefore be used in conjunction with other forms of identification, such as a passport or drivers licence. When it is used to see how much tax you owe the government however, there is a high level of confidence that the person using the number is the correct one. However it is important to distinguish between these two situations, as the first instance is of someone trying to claim benefit, i.e. a job, whereas the second is an instance of claiming a liability. The key difference here is that the claimant gains nothing from being taxed and loses nothing from getting a job, but as to which one is therefore more likely to be defrauded, that can be less clear cut. Tax evasion can be achieved if the claimant gets the job using a false National Insurance number, so both are intrinsically linked. So authentication in these two instances will be different, because the government has provided the card and the number and so it is a liability claim when the amount of tax owed is checked, whereas when the number is used to get a job, then it is a benefit from the number the government has provided.
The method of authentication in the case of a liability claim is to compare the database of National Insurance numbers with the number registered with the place of work. The key point of note here is that the authentication reliability is only high because of the authentication methods used in order to obtain a National Insurance number. Many records and documents are required to procure a number, which is also used by employers to verify that a potential new recruit has the right to work in this country. This means any authentication reliability of the National Insurance number is dependent on the reliability of the former. Hence, claiming a benefit, such as a job requires the additional authentication.
Voice recognition has a much wider scope of authentication methods, as it is much more widely used. Authentication of a voice happens within the mind of most, as previously stated, whereby a person instinctively makes a judgement about whether or not a voice does in fact belong to the person being identified. The reliability of this method of authentication varies hugely depending on whether or not the claimant is visible to the verifier. However, this rarely matters, the exception being in situations involving a third party such as a judge in a court.
{{notes access="MemberRazTalukdar" text="The purpose of above paragraph is not clear. How does scope relate to visibility and how does the example with a judge relate to either point?"}}
There are however, more scientific ways of authenticating voice. Computers can be used to recognise voices, sometimes more precisely than human beings. These use specifically developed algorithms to recognise a voice based on a data store of voiceprints set up using a wide range of standard words or to recognise a specific phrase that has been stored by the ‘real person’, hence adding another method of authentication. The latter is much simpler and can be used for unlocking doors and accessing resources. The more complex of the two is already in use as previously stated, in the intelligence community to find and identify callers on the phone, as well as in recorded conversations. These practical methods of authentication have both been highlighted many times in films and TV shows, such as Goldeneye and 24 in many different ways, which shows the enormous potential of the technology. Once the task of authenticating a voice has been completed, there are numerous applications, such as voice control, which can already be seen in devices such as mobile phones and Apple Macs.
The authentication methods have advanced greatly for both identification methods since their inceptions. Voice recognition has been around for as long as man, but the more scientific methods only became widespread after computers, because of the advanced nature of the algorithms. Experiments in voice recognition started in the 1930s and since then it has grown with increasingly powerful computers (McSwain, 2007). Now there are many different technologies to store process voiceprints, which use various different techniques to compare the voiceprint to a claim. Frequency estimation, Matrix representation and pattern matching algorithms have all developed over the years and the ever-increasing complexity of voiceprint processing is enabling more and more applications for the new technology (Speaker recognition, 2007).
National Insurance numbers have only been changed once since its inception in 1911, when it was expanded in 1946 (National Insurance, 2007); however the methods of obtaining one have changed and are becoming ever more stringent. Although every child born in the UK is assigned one automatically, foreign citizens coming into the UK must pass a number of tests to obtain one. In 2006, the Home Office outlined its Immigration plans, which from 2008 include the Biometric identity being established before a National Insurance number is issued (Byrne & Reid, 2006, p.52). Suffice to say, the National Insurance number is unlikely to change in and of itself, but the authentication methods are always changing. However unlike Voice Recognition, the authentication methods use for National Insurance numbers are often politically motivated rather than being improved because of scientific necessity.
The question of who benefits most from these identification methods comes down to where they came from in the first place. While some may argue that UK citizens benefit most from the National Insurance number, which allows them to use Public Health Services, receive benefits, etc and considering that the government originally created the National Insurance number to be separate from general taxation (National Insurance, 2007), over the years the barriers between National Insurance Contributions and Benefits have been weakened. This means that the public now directly fund Benefits, saving the government time and money. Add this to the fact that the Number is now mostly used when getting a job and therefore for taxation, the government is the clear winner here.
With Voice Recognition, the potential of the technology means that there will be many beneficiaries from this method of identification. However, applications such as voice control are still in the early stages and not that widely used. The Intelligence community is already employing voice recognition wholesale and use it for phone taps, identifying people in a reordered conversation and even for identifying agents on the phone using computers. The military and defence departments of many governments always get the biggest budgets and it is therefore natural for these parties be the first ones to employ the technology, because the costs of such advanced voice recognition systems is huge, due to their complexity and need for vast amounts of computational power. As it stands, there is no one else using voice recognition as a key part of their strategy for business, commercialisation or security. So here too, it is the government, or at least a part of it, which is drawing the most benefit from this method of identification.
The debate rages on about rights, security and privacy, but what these issues? Collection of personal data is often required for identification and the government continuously asks for updates to their data stores on us. Many people dislike the idea of giving out personal data; however it is often compulsory to do so. In the case of National Insurance numbers this is not really an issue, as the government already has the data and would collect and store it even without the number. In fact, the National Insurance number protects the personal data of individuals, as only the number need be given out to companies, employers, etc. The government does the check necessary to ascertain the identity of an individual from the number using their own data stores.
Voice Recognition on the other hand can be completely neutral in this regard. It can require only a voiceprint to identify someone if it is a simple routing system for a call centre for example. However, in the context of intelligence applications such as phone taps it can be the exact opposite. Tapping phones are seen as high invasive and even against human rights, with many people actively trying to stop this kind of invasion of privacy in their countries. Phone tapping is controversial that evidence gathered through this method is not admissible in a court of law. So while one method of identification helps protect anonymity and privacy, the other can be used to actively destroy it.
Identification methods are many and varied, but the two explored in this article represent almost all of the attributes used in all of them. Numbers are often used when it is essential to have uniformity, as well as uniqueness and are usually linked to a collection of information on a given individual. Natural attributes, such as voice, are easier to pass off for authentic identification with the aid of computers and are often inaccurate (As a small demonstration, here is an audio file which took less than a minute to put together). It is a much more complex process to identify someone using their natural attributes, than it is to do so using a clear and uniform number. No single method of identification is the best overall, but instead each is best suited to different situations. Voice Recognition is an extremely powerful tool when it comes to identifying people in spontaneous situations, like a conversation in a room or on a phone when only one piece of information is available to you from one source, whereas the National Insurance number is best suited to more standard tasks, such as checking up on an individual’s earnings that can be processed over time, with as much information as is needed.