Beyond recognition

Posted on 19 Jan 2007 at 16:31

 

Gallery

Calm before the storm

If they could be made more accurate, the applications for calm identification techniques such as these are potentially wide-ranging. In a conversation, you don't usually stand up, announce who you are and begin to say your piece. The interplay is far more dynamic and informal. Writeprints could facilitate interactions between individuals that are just as natural by giving us the ability to recognise each other from what we type and also spot impostors regardless of their usernames.

"The use of [writeprints] specifically tailored towards multilingual online messages and automatic recognition mechanisms makes our proposed visualisations feasible for identification of online messages," said Abbasi. The implication is that writeprints work regardless of the language used. They should work as well using Arabic as they do English. Being able to identify writeprints automatically in any language may help prevent online crime. However, there might be serious consequences for internet users in other countries.

While calm-computing analysis might serve to protect the individual, there's always the possibility that these techniques could end up being used as tools of repression. In an authoritarian state, being able to identify and track a single user out of millions, regardless of their attempts to defy detection, could be used to the detriment of personal freedom. Maybe the most sinister aspect of these techniques is that, by being 'calm', we may never even realise they're taking place. Your mobile phone might one day inform on who's carrying it, as well as its location, as a matter of course.

Identity parade

For now, we have nothing to fear from these advances. If they do come into general use, they could be far more accurate than current biometric identification techniques. These currently all have lower accuracies than most people suspect. In a 2004 Home Office report on the user experience of biometric scanning, the figures for successful identifications were surprisingly low. While iris scans have an upper success rate of 96 per cent, this falls to 81 per cent for fingerprint scanning and just 69 per cent for facial recognition. A calm alternative could do away with all three, replacing them with 100 per cent accuracy, all without you ever knowing they were being used.

Author: Jon Thompson