Affective Computing for more effective AI solutions
I was wonderfully surprised when a colleague of mine offered to write an article on Affective Computing for an AI publication WorldTribe is carrying out for our contractors. Then I thought, well: Many people aren't yet familiar with the basic idea. It would be worth explaining a couple of words on our blog.
So, what is it?
It is exiting! But this does not answer your question. As you can see in the compound word, it is all about affection and computing. Affection in the sense of generating human like emotions. This is what you see in Sci Fi movies, where beautiful robots flirt with their male user while they fall in love with their AI because they can't tell human and machine apart that easily. Well, that at least is an extreme idea of what one day would be possible.
What does it do?
Developers try to programm kind of human-computer interfaces that "understand" the emotional state of an end user and are able to react appropriately to it. In theory it would be possible for a program to recognize a student's negative mood in class and maybe then offer extra exercises on the learned content. It therefore uses different human signals, such as facial expressions, posture, gestures, muscle tension, speech patterns, heart rate & body temperature or pupil dilation.
Why would we want that?
Wow! The potential is enormous. As critical as the usage of AI is seen by many nowadays (which is surely more than justified when it comes to matters such as equal chances), there also lies a lot of potential, hope and ideas to problem solving solutions in further development of AI systems. But for our solutions to be able to make fair and truly smart decisions, we need more emotions in technology.
I will keep you posted about further developments and when the name of our new author can be revealed!