AI and the Future of Music

cover
1 Jul 2024

Since the dawn of humanity music has played a great rolle in our lives. Imagine the tribal gatherings around the fire singing and dancing to the drums. Fast forward to our modern age of going to concerts, music clubs or sharing playlists among each others.

I once read that to our brains, listening to music is like solving a puzzle. That our neurons are trying to predict the next note and beat providing us with the satisfaction of a good guess, or the surprise of the unexpected pleasant turn.

We enjoy going to concerts where songs that we already know will be played and let the band surprise us with slight variations of the tunes we love. Classical music and jazz are probably at the extreme, where each director and every band might interpret a known song in a slightly different way.

I was talking to my kids about remixes, i.e., when DJs or other musicians take a shot at an old song by changing the rhythm or singing it with their voice. For example, the voice of Shakira singing the AC/DC rock anthem of Back in Black or Celine Dion and Anastacia singing You Shook Me All Night Long or the remix of When We Were Young by David Guetta.

We all have favorite singers, or favorite styles. Now we also have the tools that would allow us to combine all of this and create the ultimative customized playlist. A playlist were we could select a tune, whose voices do we want, and on which style. Too slow for you? Add some techno. Too fast? Turn on the blues filter...


What can AI do for the future of music?

I remember having dinner with my friend Beda some 20 years ago. Those were the times when everyone was digitalizing their music collection, way before streaming services. I remember him saying that he wanted to have a "Radio Beda" where the songs he liked would be playing nonstop. I then went to work on digital music and saw the revolution unfold before my eyes. Internet radio stations went digital, and DRM algorithms were pushed into the stereos until Spotify and Bluetooth speakers came in like a Handgranate and created a new music ecosystem.

The music labels are already suing AI startups for using their content without their permission while singers like Grimes is offering the rights to use her voice on any future song with the condition that they share the profits.

I do not know how this all will play out, but I can dream of what a future ultimate customized playlist will do:

  • Prompters: Will come with an idea of a given song that they want to be sang by somebody else, or played in a different style. AI algorithms will generate this and make the end result available to all subscribers of a streaming service. They will be paid a small percentage of the royalties.
  • Singers: Will license their voices to the AI algorithms to be used. Note that these singers do not need to be already popular. Regardless of how raspy or sweet the voice is, how many octaves the pitch, or even the style, they could be selected by a prompter. They will also be paid a percentage of the royalties.
  • Music Authors: Will license their songs for the AI to modify. This could also mean changing some of the lyrics, coming up with new versions of say, Bohemian Rhapsody. Or even mixing two or more song lyrics into one. They will of course receive royalties.
  • Music Labels: Will, of course, license their existing catalog. Yet note that while in the beginning, the big hits might come from their catalogs, in the future, it does not need to be like that. They and their lawyers will also enrich their coffers with part of the royalties.
  • Streaming Services: This will enable the creation of new songs and distribute them among their subscribers. They will be the ones cutting the cake and keeping part of the profits.

I have not seen my friend Beda in a while, but I am sure he is enjoying his Radio Beda in one of the multiple streaming services available. If you were to ask me who would nail this new ecosystem, my answer would be that it will probably be the one you least expect it. After all, during the streaming-ization of music, the ones who were the first were not the ones who made it in the end...