Is affective computing the key to true artificial intelligence?
Affective computing is a branch of computer science that merges psychology, neuroscience, sociology, and cognitive science. It is another term for artificial emotional intelligence, which refers to the study and development of systems that are able to recognise, understand and even simulate human feelings and emotions. In psychology, feelings and emotions are sometimes known as ‘affects’ – hence the name ‘affective computing’.
The concept of affective computing was first formally described by Rosalind Picard in a 1995 paper which, two years later, was followed by a book. The book considered the role of emotion within intelligence, and the possible effects of emotion recognition by computers. The general premise was that if computers are to be genuinely intelligent, then they must accurately perceive human ‘affects’ and even be able to express them. This would make them far more useful to humanity in real world situations in which reading emotion is key to interaction. The interesting question is whether expressing and having emotions are the same thing. As such, affective computing ties into debate over machine intelligence and machine ethics. If a computer is to be emotionally intelligent, it should also be fundamentally benevolent.
Ethical considerations aside, affective computing research has led to advances in communication via sensors and algorithms, as well as new techniques to assess mood through interaction and conversation. It has also led to a better understanding of how these aspects can influence personal health, and encouraged the creation of intelligent machines that base interactions on reading emotions.
Find out more in our free weekly newsletter.