Well on one side I am worried that this will be applied to political speech or other "public figure embarassements". Just file for copyright infringement and snap! Your video of the Prime Minister insulting [women|immigrants|judges] might be taken down (I'm not talking about civilized countries here, but places like this: http://journal.dedasys.com/2010/02/24/google-execs-convicted ). On the other hand, do you think this would give a huge boost to research towards sub-optimal voice and "conversation topic" recognition? For example, suppose you record your lecturer talking about Hegel. You upload the video to youtube and it gets "recognized" by matching speech to their library. It can be tagged automatically, etc. Furthermore, the other parts of the conversation can become a new baseline to match philosophy videos, and hence grow a "semantic" base. That would be even a nicer feat.