The BBC is reporting that a Brazilian judge has ordered YouTube to remove a clip of supermodel Daniela Cicarelli romping in the sea in Spain (ya gotta love the word ‘romping’). The problem is that the initial order was issued in September and the file has since appeared under various guises. The court is ordering that YouTube do more to see that it doesn’t appear again but how are they meant to scan for something like this?
To my knowledge, it would be very difficult for an algorithm to detect a particular video. Perhaps you could do a check on the file size but thousands of videos would have similiar if not exactly the same sizes. Of course the video descriptions and tags could be checked but will that be 100% successful? This all follows a call early in December from a Japanese rights holders group calling for YouTube to proactively check if videos infringe copyright. I can see YouTube needing more staff over the next few months.