Two stories about music AI that jumped out at me this week

Hey there!

I am a staunch critic of music AI. For the most part, I think that the more any producer depends on it, the worse they will be at their craft. Full stop. I could give countless reasons as to why I think that's the case, though it mainly boils down to the fact that growth requires effort. AI robs you of growth. But, that isn't even my biggest gripe with AI.

My biggest gripe with AI is how it is structured in a way that makes it so clear that artists don't matter and that they are ultimately just a resource to be mined. This is the sticky, behind the scenes stuff that points to either the strength of the music community or the rot in our culture. Two stories published by MusicTech this week highlight each of those points vividly for me.

First up, this story about Diplo advising everybody to use AI or "give up and become an Uber drive." Right off the bat, my sympathies to any rideshare drivers that have to cart Diplo around. It seems pretty clear that he doesn't respect you or, at the very least, holds himself in much higher esteem.

But the quote from the article that struck me the most was this, “I’ve had some voices that I’ve made with AI, and I’m like, ‘damn, I couldn’t even get this take out of the best singer,’” the producer says." It's difficult for me to not read that as anything other than: "AI is better than a real human collaborator because I can extract more from the algorithm than I can from the person." It makes you wonder, what's the point? The answer is, of course, to make a bigger hit to make more money and to have more fame. AI music is about anything other than the music. It's cultural rot.

On the complete other end of the spectrum, is this article about Splice rolling out new AI tools that compensate creators. Each of these tools essentially edits a sample in a particular way for the producer who wants to use it and the tools ensures that the source sample creator is compensated. This is absolutely a step in the right direction. This leans more towards assistive AI, which I find tends to be the kind of AI that has a better chance of supporting the creative process and not usurping it entirely.

I still have questions about this new Splice tech. What about the samples used in the training sets for this algorithm? Were those artists aware that their work was being used as training data? Are they being compensated? Does this process happen locally on the producer's laptop or does it require use of a monstrous data center? I will repeat that Splice's new model is a step in the right direction but there is still plenty of climbing to be done.

We're getting inundated with AI on all fronts and it's more important than ever that we think critically about the role these algorithms play in our lives and careers.

What do you think about these stories?

Til next time.