[JohnnyWaffles]

So I thought this might be an interesting thought  to post. I'm sure its been discussed before, but consider for a moment your favorite music, and how it makes you feel. There is no doubt that music has a profound affect on our mood, if even for a moment. It can stir and create emotion in an individual. I know myself, if I have certain music on when I'm driving it pushes me to drive aggressively, which isn't always a good thing  . That's just one example.

I'm curious if you could use sentiment analyses to try and emulate the affects of music and audio on an AI. For training data you would you need a way to track "emotions" in people when listening to different kinds of music. Outside of using computer vision and facial expressions I'm not quite sure how you would be able to track this. Plus a facial expression may not mean anything to an AI in terms of emotion in the way it would if say, a human being saw someone with a sad face. We are social animals that are pretty good at empathizing with people and detecting emotion.

I wonder if this is one approach to creating "emotion" within an AI. The question is though, do you really want emotion in an AI? Emotions help to make the human experience really, and it's what helps to create a satisfying existence. If you never experience sadness, you will not know the quality of happiness.

One advantage I see to this would be AI to Human interaction. If it could "sense" the way we feel, it would only improve it's interactions with people.

[HS]

You could use biometrics and brain MRI and camera's, to get a detailed emotional map. But being observed would change the reactions. Just look at reaction videos on youtube  ... You would have to do prolonged non invasive detailed monitoring until a person forgets that they are being observed.  So that part seems doable. If everything went well, then you'd end up with an emotion parrot. Which would be really cool, but it seems to me that it wouldn't be alive until it got some recursive internal stuff going on.