Biometric flaws mar start of ID card plan

The government has taken the first steps in the technology procurement for its national biometric identity cards programme.

But experts are already questioning the maturity of the biometric systems at the heart of the proposals.

The legislation required for the scheme is not yet law, but the Home Office last week issued the first notice alerting potential suppliers, to ensure the procurement can start as soon as the bill is passed.

The formal tender will include the creation of a national identity register, biometric technology, and the production and distribution of cards.

The scheme will create a gold standard of identification to help fight fraud, terrorism and illegal immigration, according to the government.

But biometric technology is not accurate enough to support such a claim, says Neil Fisher, director of security solutions at defence technology supplier QinetiQ.

'Telecoms systems are judged on an availability of 99.999 per cent, but even that level of accuracy of biometrics, across the whole population, would mean 6,000 people in the country being mistaken, and no biometric technology is anywhere close to that reliable yet,' Fisher told Computing.

'Unless there is a strategy to overcome that lack of accuracy, the system will be flawed as soon as it starts,' he added.

Plans to include multiple biometrics - two eyes, 10 fingers and one face - offer some improvement, but checking all 13 will be a long and expensive process, says Graham Titterington, principal analyst at Ovum.

And the belief that biometrics mean no one will be able to register twice is incorrect.

'At the moment, the whole reliability of biometrics is up in the air. There have been lots of studies done with very different results,' said Titterington.

'The government needs a dose of reality because its trust in the system is unfounded and doesn't match up with experience.

'The plan is working on the assumption that, by the time it is live, the technology will have come on in leaps and bounds. But that is not a reasonable basis from which to start.'