Why do we need affective computing?

One practical example: Intelligent Tutoring Systems are a meaningful domain for application of Affective Computing. Imagine for a moment a learning system that can tell when you are frustrated or angry, and adapt the training scenario based on mood and performance. Training efficiency can improve, which is a potential for real world resource savings.

There are aspects of many important psychological phenomena that can be better or only studied using affective computing. In my own work, I have used automated facial expression coding to gain empirical insight into depression. Future work will do the same with personality, group formation, marital and parent-child interaction, etc.

It's not just about making virtual humans or computers that imitate humans. Much do of what humans do with computers involves emotion in one way or  another, or might in the future be improved by taking it into account.  Many current examples are primitive and/or hokey.  But the methods used in, say, a toy human (non-human player character in a game) would find uses in other applications (tutoring, animated dramas, recommendation systems, personal shoppers, content quality bots .. you name it).   I think we are far from understanding the implications of this change,  both good and bad, and what we need to think about to manage it.  A common vision is that humans and their machines will blend, creating "thick" technical environments that reflect and respond to our feelings, and more importantly, perhaps, intentions, motivations, and desires. For more, click on over to for more answers and visit http://en.wikipedia.org/wiki/Affective_computing on Wikipedia.