Professor Luis von Ahn at CMU has created Gwap, a research tool in a similar vein as Google Image Labeler. It’s actually a fun diversion from the day to day. In the guise of a game, Gwap is a little more rich than Google Image Labeler in its offering. It allows the user to:

-map words to images
-come to a consensus as to whether you are listening to the same song as your partner
-map words in sentences to help describe other words
-trace aspects of an image based on an image and a word. the closer you and your partner trace the image similarly the more points you are rewarded.
-pick one image of two you believe your partner will favor

As we gather more human data and allow machines to process that information we can hope for a more friendly semantically aware web.