Rick Beato certainly has the evidence:
Yes, a complete neophyte can sit down with some AI programs and in a couple of minutes "create" a new, superficially plausible pop song. What can we draw from this? As Rick says, this is no substitute from actually learning how to play an instrument and, probably more important, learning how to create, invent, discover (whatever the appropriate verb is) music. I've been ranting against computer derived music for years, but this is a kind of nadir. My view for a long time is that much pop music, at least the commercially successful stuff, is little more than an industrial product. As Rick comments in a different video, no, pop stars do not even write the lyrics to their songs, so they are in no way a personal expression. I rather doubt that most pop music could even be considered an aesthetic object. It's more of an acoustic equivalent to valium or, in the case of heavy metal, amphetamines.
Let's get that stuff out of our ears by listening to some Bach.
AI pop is why music professors STILL have to keep assigning Adorno as required reading in 2025. But Adorno, of course, would've said most of the rock that Rick Beato likes might be just as dumb and formulaic. On the other hand, I can't forget that passage in Current of Music where Adorno took the trouble to say that "Deep Purple" is alright for a pop song and "Two in Love" is an obvious, lame knock-off in terms of gestures and phrases. Adorno had moments where he'd pull a Roger Ebert and say "You shouldn't make your diet entirely hot dogs but, hey, if you want a hot dog this kind is good".
ReplyDeleteI am suddenly reminded that there are, this very day, a couple of thousand subscribers to the Emily Howell Youtube channel
ReplyDeletehttps://www.youtube.com/@emilyhowell2359
If AI made subpar pop songs in 2025 know that it was used to generate subpart classical music years earlier. :) If AI could make bad Bach as far back as 2009.