By Heather Hayes (with additional BLUE color comments by me)
June 30, 2003

Which biometric to use- (June 30, 2003)

E-government applications that use biometrics to authenticate the person making a transaction are likely to rely on one of four types of identification: fingerprints, facial recognition, iris scan or voiceprints. Choosing the right biometric for the job could mean the difference between success and failure.

All the biometric choices have pros and cons in an e-government environment.

Because it relies on a telephone, voiceprint identification doesn’t require any special equipment or training for the user. It’s cost-effective for both the customer and the agency and is often seen by users as the least invasive biometric technology. But there are downsides. If the telephone connection isn’t clear or the user has a cold, a voiceprint can be difficult to identify.
Or if someone records your voice …

Fingerprint readers are now being integrated into keyboards, and stand-alone units are dropping in cost, making them easy enough to use in the privacy of one’s home. And they are especially well-suited to government- to-government and government-to- employee applications, because agencies typically already have federal employee fingerprints on file. On the other hand, the criminal connotation associated with fingerprints can make this method a hard sell with the general public.
Gummi bears are one way around this.  Disney has been collecting fingerprints from theme park visitors since 2005.

Iris scans.
This is by far the most accurate of the biometric choices, but it is also the most invasive and expensive.
I’ve seen this in use.  It was used to protect one (Windows NT4 PC), which reportedly had confidential, BUT not classified information on it.  The reason for the particular agency buying it?  The CIO was a gadget freak.

Facial recognition.
Users can rely on their own Web cameras as readers, making it more cost-effective than other technologies, but it must still be perfected to reduce the number of false positives and false negatives.
Only now starting to see wider deployment of this.