According to an article on Wired.com, collecting the information that allows a filter to compare an image, a book, a song, or a video to existing material is itself an uncompensated use of the material it is meant to single out for violation. In other words, even “fair use” isn’t enough of a reason for website to police themselves prior to actual notification by a copyright holder since the website would have to have legal copies (aka paid for) of everything being filter to compare it to.
If this holds up, no website can currently be held responsible to filter its own bandwidth for violations because they would have to purchase and own a copy of whatever is being searched for (read: everything and anything) to check and see if it’s real. Ha!
ThinkingSkull.com 