If there’s anything we can be sure of in the sphere of music is that there is always a counter-reaction. We’ve seen this in genre development through the decades. As one genre grabs the spotlight, another genre brews and grows in the undercurrents. It’s as if artists instinctively provide a balance of sounds.
This is also true when it comes to methods of creating music. Overly produced approaches create a demand for raw, authentic sounds. These underground reactions take time to evolve before they bubble to the surface. Then as these sounds become commercialized, the genre solidifies and in time becomes stale.
In 2023 we’re going to hear more and more about AI-generated music. Technology is developing, and we already have a tool to do this: Mubert. This “text to track” AI tool generates music from over one million samples from 4,000 artists.
Using Mubert Studio, Artists provide sample packs. Artists upload several variations of each instrument and submit them as bass, midi, pads, leads, perks, drums, vocals, claps, hats, or fx. Loops can be submitted, from one to 64 bars each. Artists receive royalties each time Mubert uses their samples or loops (I have yet to see a detailed look at how this works).
We suspect there will be a flood of similar AI music generation tools, some of which won’t be as transparent about where they draw their samples from. We can be sure that it won’t be in the best interest of many creators. And it won’t be long before there is a flood of AI-generated tracks added to Spotify and Apple Music. This will add to the already crowded expansion of music submissions, which is currently over 100,000 track submissions every day.
Going back to the reactionary tendencies in music, I expect that we’ll see a growing number of underground genres that emphasize authenticity, true artistry, and genuine music creation. With the growing number of home studios, we can expect more growth of artists creating in their bedroom (figurative and literal).
Personally, I’m excited to hear what artists create in their homes, rather than text-to-track meshing. (Below is an image I created on DALL-E from the text “music generated by computer.”)
