This AI Is So Human That It Can Detect Emotions In Music

IBM’s Watson AI is now a music psychologist.

If you were already unsettled about the idea of artificial intelligence making music, then the idea that IBM’s Watson A.I. can breakdown musicians’ psyches by listening to their music must be pretty terrifying. Its “Tone Analyzer” software was tasked with running through thousands of songs to determine their overall mood and emotional content. It broke down each track and categorized them by their levels of anger, disgust, joy, fear or sadness (you know, because some of the most memorable songs in pop history are shallow like that).

Putting the glaring methodological issues aside for one second, the results are actually slightly entertaining, if not interesting. For example, it seems Elvis was a pretty angry guy, an emotion only outdone by his absolute joy which is close to 100%. Rihanna’s songs barely have any emotional content bar fear, while Adele’s sadness seems to be her defining trait. Nirvana on the other hand was just misunderstood: We all thought Kurt Cobain was angry, when really he was disgusted all along. Check out IBM’s detailed infographic on the results here.

Read more: Google’s new AI can use theory to make better music than you

(Via Hypebot)