Monday, March 21, 2005

LSE publishes draft paper on ID cards

The LSE has just published a preliminary report on ID cards. It doesn't like the gov'ts current proposals much. A few things that caught my eye on the way through.

It looks like being a major imposition, with the biometrics needing updating "every five years for much of the poulation." It will also hit the disabled and elderly harder because their biometric tend to change at a greter rate, such that they may have trouble proving their identity against the biometric referenced by the card.

There is nobody in charge of ensuring that the data stored is accurate. There is no liability for inaccurate data.

The gov't arguments regarding international obligations for biometrics are destroyed. Even in the US which currently has the highest biometric requirement, the biometric is a digitally stored photograph (!?!).

Because refusal to register is a civil liability not a criminal case, wealthy people and wealthy criminal gangs can avoid ID cards by paying the repeated £2500 fines.

A central ID number with access to multiple services is very valuble to criminal elements. The more important a single item is, the more it will be targetted. In a number of countries where a cenral identifying number has been implemented ID fraud has increased.

Benefit fraud through multiple IDs accounts for £50million per year. The ID card scheme will cost £5.5billion over ten years (note Billion)

"All biometrics have successfully been spoofed or attacked by researchers."

"The vast majority of biometric trials have been in the 'frequent traveller' context, using volunteers who are predominantly white male professionals in the age group between 20-55 years old." In other words fit, healthy people who present no major problems for biometrics. A well tested system then. The trial featuring a more diverse group has not been published.

The alternative proposal on page 71 is particularly interesting for nerds who can understand it. As part of the spec, the report specifically accuses the government of wanting a full disproportionate flow of information about all subjects, rather than allowing subjects to control their own identity. A serious charge indeed.



Chameleon said...

Thank you for an excellent link to the LSE report.

Scribe said...

Hmm, really must put together a page detailing links for all these reports...

I thought I read something about why digital photos are used, although it may have been a dream. Off the top of my head, I would have thought that it was because a). face recognition has plenty of "promise" left, and b). it's very easy to verify manually when the photo is printed out (compared to, say, retinal data).