I fully acknowledge that this is the kind of thing that can give privacy advocates nightmares about what Facebook and Google are capable of, but I'm incredibly excited by ideas like this. Things like this are what keep me interested in Computer Science, the possibility of being able to capture and study real, raw data from people outside of a laboratory setting as they respond to the most important current events of their day. Again, there's a lot of potential for abuse, but I'd be intrigued to see what lawmakers and corporations could do with information on how people at large "really" felt about them.
As someone who is studying Computer Science, I understand and even share your enthusiasm for exciting new ideas. The technical challenge that a problem like this presents is very, very interesting. That being said, history is full of exciting scientific problems whose solutions have been abused. I certainly don't have a solution, but it seems to be a common theme associated with scientific innovation :-).