A cybernetic Bjork in the “All is Full of Love” video, 1997

Robot Rock: Will Artificial Intelligence Replace Musicians?

New music creation technology is trading artists for algorithms


Take a look at any of the “cab drivers versus Uber” protests that flare up occasionally, and you’ll see a perfect example of short-term thinking without the benefit of a long-term perspective. Sure, Uber might be displacing cab drivers in right now — but what happens to everyone when self-driving cars hit the market in ten years or so? Rather than focusing on hedging against that possibility, it becomes easier to fight the enemy that is immediately in front of you — but eventually, everyone is going to lose.

This popped into my mind when I saw that a company called Jukedeck won the Startup Battlefield at Techcrunch Disrupt London late last year. Jukedeck allows users to fill in a few items (genre, mood, length of track) and then get a custom song made just for them by an algorithm. The site is initially marketed at people who need background music for videos or small businesses who can’t afford to license existing music, and co-founder and COO Patrick Stobbs told me it isn’t meant replace human musicians — if anything, it can make their lives easier by doing some of the composition work for them.

Experimenting on the site, I created a few tracks and can say that while I’d be happy to use them to soundtrack a vacation slideshow or a video of my dog, they won’t be topping the Billboard Hot 100 any time soon. But we’re also still in the early days of the product and the technology, and who’s to say how it will grow in the next few years?

We’ve already seen massive disruption in the music space at the hands of the machines. What used to take days in a studio now takes a few hours on a laptop, and many more people can participate than before. One of the reasons Soundcloud has such a huge user base is that it turns out people really enjoy hearing remixes done by amateurs and friends, not just “experts.” Anyone, in almost any part of the world (assuming they have internet access) can create a song and share it globally with just a few clicks. We’re also starting to see more programs that allow people with almost zero musical knowledge to make music, such as Looplabs, which I called “Minecraft for Music.”

All this said, I don’t think Justin Bieber needs to worry about being replaced by a robot yet. Music is almost as much about personality as it is about the product, and despite a handful of sci-fi movies to the contrary, we really don’t have to worry about teenagers falling in love with machines quite yet. But this also doesn’t mean musicians can bury their heads in the sand and keep fighting the battles in front of them, when there are much bigger disruptors than Pandora and Spotify looming on the horizon.

First of all, think of all the times we listen to music throughout the day, and how for much of that time, the music is just background noise. I’m technically listening to music every time I go to the grocery store, or drugstore, or a bar — but I’m not really hearing it. Rather than having to deal with licensing, or collecting bodies, or playlist makers, it would be much easier for public spaces to just type some keywords into a service like Jukedeck and let a machine create the music to pipe in while shoppers browsed for cheese or soap. TV, film, and video producers could also do something similar — it’s cheaper and solves the problem of hearing the same song in different shows, to the point where it because a joke (I’m looking at you, “Hallelujah,” played during every sad moment ever on TV).

There are plenty more examples where music generated by an algorithm could work in more intentional listening situations. Take working out — instead of slogging through a run while listening to another round of remixes, a program could create something fresh and custom every day, based on distance and pace. You’d never have the problem of the same song coming up on Pandora or a song on a Spotify playlist that throws everything off. If you need music for working, or commuting, or really anything aside from full on listening, it could probably be created by a program in the next few years.

For artists, this could shrink the pie dramatically. As I said above, fans are drawn to personalities as well as songs, and for artists with strong personal brands, there is still money to be made on music and touring. But for artists that are less interesting and don’t have strong live shows, this could pose a real challenge. If sync licensing income dries up and streaming services add AI options and capability, resulting in fewer streams, income could disappear.

Stobbs says that Jukedeck is democratizing who can create music, and making it easier for people with no training to explore songcraft. Much like YouTube has done for video stars, Jukedeck and Looplabs allow everyone to compose and share something with a niche group, and perhaps make it big. For professional musicians, this is a development worth watching and potentially preparing for — looking the other way as the machines rise is a surefire way to be run over quickly.


If you enjoyed reading this, please login and click “Recommend” below.
This will help to share the story with others.

Now you can buy whole book of pieces like this, along with long interviews with some very smart people. “How We Listen Now: Essays and Conversations About Music and Technology” by Cortney Harding is now available via Amazon.

Follow Cuepoint:Twitter|Facebook

Show your support

Clapping shows how much you appreciated Cortney Harding’s story.