Here's my take. Within music production I like to consider two worlds of sculpting sound. Analog clay, and digital clay. Analog clay is any physical instrument you can record and later shape in the box. Digital clay can be done with software. However when it comes to software, to grab a ball of digital clay and shape a source from it's primitive is either going to come from a stock instrument, a sample library, a soft synth or anything that can be edited in midi. To consider ai digital clay is simply not true because you are not getting source material. For example, instead of manually processing the signal of a dry sine wave, with ai, you're generating a random output. Or maybe several outputs until you're satisfied. Which is fine, but personally I don't find it enjoyable to render a prompt until I'm satisfied with the output but to each their own. I love the process of creating music, I love shaping my digital clay.