Wear Dark Glasses

Regan writes "The December issue of Scientific American has a scary story on the use of facial recognition software to identify "subversives" among the general population. I did a PhD in image recognition and have serious doubts about the reliability of this software. Automatic face recognition is hard, even when the subject cooperates. What are they going to do about false positive identifications? If you're suddenly bundled into a car the next time you're walking past the post office you'll know the answer. " Interesting point: What do you folks think? Can this work?