• silicon.com
• Management
• Public Sector

By Brian White, 16 January 2006 15:15

COMMENT

A national ID card may not be bad in theory, says Brian White but there are a number of technological kinks that need to be worked out first for it to succeed.

Most of us are familiar with the 80:20 rule, whereby 80 per cent of the work in any project can be done in the first 20 per cent of the time available.

It's the last 20 per cent which causes all the problems. And in the public sector you only need one issue to undermine the credibility of the whole scheme.

The government's ID card scheme is a classic example of the problems that need to be addressed when a universal service is required. Most private sector projects can tailor the system to key objectives and groups but when every citizen needs to be identified you start to get a number of intractable issues.

After all, how can you identify the terrorist if you have a system which can only identify 95 per cent of the population?

We know that biometric testing is not an exact science and there are a number of people who come up as false positives. Even the best type - DNA testing - can only go as far as saying the chances of this individual not being the right person are statistically remote. It cannot say with 100 per cent certainty that it is the same person.

So people will fail a biometric test. The iris changes as we get older, and there is something like one in 70,000 people whose iris cannot be machine read. During testing there was even a whole ethnic group whose iris was not readable.

It is inevitable that some cards will fail the test so the question is how are the authorities going to deal with these people? Will they be treated like criminals or potential suspects?

The government is faced with a real dilemma in how they will manage this process.

If the government sorts out all the problems and achieves - say - a 99.7 per cent success, then we are still left with 135,000 people failing the test and having to continually prove that they are who they say they are.

And if the pessimists are right then that figure will rise dramatically.

Then there is the question of who will do the testing and where. Will we have roadside ID card checks? Will the private sector have access to the technology? At what level will these testers be in their organisation? And what analysis has been done into how the use of this technology will affect the way organisations operate?

One US company which introduced iris recognition technology correctly estimated the time it would take a reader to scan the person - but failed to realise that each person would waste several more seconds preparing themselves for the reader by straightening their tie, combing their hair and generally smartening themselves up.

The system was quickly withdrawn when they simply could not get enough people through the turnstiles to be profitable.

Over-reliance on technological solutions or a belief in their infallibility will be the undoing of the ID card.

There is no doubt there are real benefits to the introduction of an ID card - which I support - but until the government deals with the social aspects of the introduction of the cards we will simply be creating another IT project which went wrong.

As always, we seem to be blinded by the technology and we fail to appreciate how people will react to its introduction. That is where the government's tax credits system came unstuck and we don't seem to have learnt any lessons.

Brian White is a business adviser and former MP for Milton Keynes North East. When an MP, he was treasurer of the Parliamentary ICT Committee and an officer of the All Party Internet Group.