AI and Skill Erosion

 

Eccles Tower Erosion

 

If you use a particular technology on a daily basis, as I have with AI music for the past two years, it begins to shape your way of working: You find it enjoyable, creative, and "generative" (in the traditional sense--where one idea is creating others), but there's continuous "background hum" of doubt about it, or is another example of "aesthetic erosion", where the shoreline of how you worked in the past is being slowly eroded. So far it hasn't affected my musical skills, but it is still a collective force, perhaps changing the creative routines of pre-internet (or even pre-computer) musicians. I still continue to work in the old ways and keep sketchbooks as a way to shore up the cliff. A part of the cognitive dissonance is that AI music is really not music, but we now have the capability of migrating the old technologies, for example, using an audio file generated from a MIDI file to be remodeled by AI. But the music still isn't music in the sense that it can be performed by humans. I can take a 3-second idea for a wind quintet, and it will spin it out to 3 minutes. But there's no score or extracted parts. AI gives us the illusion that things work out of the box--and they do, and the results are impressive, but they're still just exemplars of music. You can't just print the score full of thousands of errors and tell the bassoonist to "just read the bottom line". My fear is that too much use of AI will result in musician brain rot, including all the things that music skills might be connected to in the brain, and all that is going to get pruned out over time. But that has always been the case. At least in the 60 through the 90s, people were playing instruments., The internet had not yet become an instrument. 

Comments

Popular Posts