Can a System Be Intelligent if It Never Gives a Damn?

Proceedings of the AAAI Conference on Artificial Intelligence, 5

Volume

Science

Cognitive Modeling and Education

I explore whether all types of cognitive faculties are possible in a system which does not also have "affective faculties" such as motives and emotions. Attention is focused on the human cognitive faculty of understanding the human affective faculties, and it is suggested that we do this in part by using ourselves as models of other people. Therefore any system which could perform as well as we do on this task would incorporate much knowledge about human affective faculties, and would probably have embedded within it a "model human" which would possess affective faculties. However, this would not necessarily manifest itself directly in the embedding system’s behavior; thus the embedding system, unlike the embedded one, need not have affective faculties. That’s a relief, because we might prefer that it didn’t.