The big difference with AI music tho is that it is fundamentally non-musicians using prior musicians' work to train machine learning algorithms to spit out dozens or hundreds of songs within the span of a few hours.
We're not talking about a difference of styles, or musical approaches or even instrumentation- we're talking about the ability of a single coder or small group of developers to build a one-time app that can then take the collective creative output of decades of musicians and begin to mass produce derivative works ad infinitum with zero effort, zero credit or recognition, and zero compensation to the musicians who produced the source material.
I'm not worried that AI art/music is going to replace bands/galleries/concerts/albums etc... I am worried that it is going to decimate session musicians, film and video game scoring, commercial jingle writers etc... and do it by stealing from the people it's trying to replace.