The technology in question starts with "gigapixel" photography. Gigapixel photos are giant panoramas that themselves consist of hundreds of component mega-pixel digital shots. This means that you can begin with, say, a distant view of London or Seville or San Francisco -- and then keep zooming in until you are looking at individual buildings, cars, street signs, people. It's the effect we've come to take for granted with Google Earth, but with much greater close-in photographic clarity.
When you combine that with better and better automated systems of matching faces to real identities, plus human-run social-network tagging systems, what do you have? The result is something familiar in dystopian sci-fi novels from 1984 onward but until now not part of real life: the impending extinction of the "faceless crowd" and removal of any mask of anonymity as we go about our daily affairs.
If you go to this site, you'll see what I mean. It starts with a big crowd scene in Vancouver, soon before the Stanley Cup riots, which were at this site and presumably involved some of these people. If you start zooming in on the photo, you can get close enough to almost any of the tens of thousands of faces to see who it is.
Here is an idea of the big picture:
And here is a zoomed-in bit from near the very back of the crowd, so far away in the shot above that the masses of people are just a blur. Up close, the people at the back look like this. You can imagine how ones nearer in look:
It's not hard to think of ways these technologies can combine for surveillance-state purposes. For a few implications, see the Daily Beast, Canada.com, and the Frankfurter Allgemeine in a Google-translated English version. Also, if you go to the Gigatag view of the photo above, you will see that, via Facebook, large number of faces in the shot have already been tagged with real identities. By their "friends."
Maybe we'll all start wearing bandanas and hoodies when out in public, following the lead of the recent rioters in London (where monitoring cameras are already everywhere)? More to think about later, but for now the site is worth trying yourself.
UPDATE: For further discussion of the ethical and practical implications of "recognition markets," like the Gigatag effort to have "friends" identify members of the Vancouver crowd, see this 2006 academic paper by Ryan Shaw. One of its conclusions:
>>"Ultimately, the best solution for protecting visual privacy in recognition markets may be education. In particular, people need to be aware of the consequences of indexing images for searchability and of linking images to other kinds of data."<<My translation: if you have a chance to identify people in crowd shots, don't.
This article available online at: