Artificial Emotion for AI

Emotion can be categorized as: happy, sad, afraid, angry, trust, disgust, surprise, anticipation. These groups are from Plutchik’s wheel of emotions. He postulated that an entire spectrum of emotional states can be derived from combinations of eight basic forms. There are other similar taxonomies of emotion.

Various emotional states are associated with physiological state changes. We also characterize emotional states as being “subconscious”, or at least as a process that occurs with less awareness than our full analytic ability. We know that our brains process a lot of information at levels below our normal conscious awareness. That alone does not make the results of subconscious processing more primitive or inaccurate.

Some analysis and predictions need to be made quickly with less information than we’d prefer. Intuition describes a predictive state that we can’t fully describe. Emotion describes a reactive state that we can’t fully describe. Intuitive analysis and emotional reactions seem to be a way for our minds to reach conclusions quickly and take action promptly, even though we have not fully collected or digested all available information.

Sometimes survival depends more upon rapid reactions and sometimes it depends more upon accurate reactions. Our fast intuitive and emotional processes can keep us alive when we need fast and our analytical reasoning ability can keep us alive when we need accurate.

Artificial intelligence has the potential to both operate faster and analyze more information than a human mind. If a problem has been correctly framed, the speed of an AI will be superior. But when a real life problem is “fuzzy” and requires a lot of pre-processing decisions to determine which information is most pertinent to a problem, some form of judgement will be needed to balance speed against depth.

A “resolution determination” might help decide what level of graininess will offer the best decision in the least amount of time. Previous experience could be used to make a best guess at that resolution level at the beginning of the analysis. This process could be considered a form of AI emotion.

SEE ALSO:
Intelligence = Ethics = Friendly AI
Emotional Awareness
Ambient Intelligence

Comments are closed.