The idea that AI can compose electronic music may sound a little off to people. It raises essential questions about creativity as a product exclusive to humans: can an AI be creative? Can it be musical? Can it compete with human-made melodies? Does it need to?

More and more, AI has set foot in the realm of creative industries. From an AI writing the next GoT book to IBM’s Watson creating a trailer for a non-exisent sci-fi thriller. And that’s not where it ends: the music industry also got involved when that same Watson was used by award winning producers to create country rap, not to mention a Eurovision song created with machine learning.

Electronic music, too, is affected by the algorithmic technologies that revolutionize the way humans relate to the arts. As a discipline that has technology at its very core, electronic music is bound to cross paths with the ways of AI. From DJing to producing and from contriving DJ names to directing music videos, algorithmic agency is growing stronger each day.

The subsequent question is how humans pertain to these technologies and how the arts and AI can be treated as a symbiosis, rather than a dystopian binary. Put differently, how can we embrace AI as an instrument we can work together with, rather than an autonomous entity overruling human creativity?

Music has always been technological

There has always been a link between music and technology, as essentially it revolves around counting and measuring the rhythm, as much as it relies on instruments.

Clapping their hands, our early ancestors used their body as an instrument to create rhythmic music. As our predecessors found out that they could smack sticks or stones to enhance the beat without hurting their hands, drums were invented.

Fast-forward to the 20th century, elaborate drum kits emerged at the intersection of African-American brass bands and western instruments. The technology of the bass paddle made it possible to use both hands and feet to incite sound, hence evolving the drum kit as we know it now.

The instrument was further technologized when companies like Korg and Roland started producing drum machines on a massive scale. The genres that emerged from these instruments diverged, but essentially, both the drum kit and drum machine serve as a technology to produce the rhythms and sounds that we know as music.

Are algorithms are the next DJs?

In the same line, DJing has undergone changes when the vinyl decks were complemented by USB-driven CDJs. Though the technologies changed, the art of DJing remains present – just in different ways.

In this day and age, AI is the upcoming technology broadening the horizon of (electronic) music. On a day-to-day basis, algorithms are already silently ruling our music taste through auto-playlists like the ones developed by YouTube, Spotify and Apple Genius. In a way, algorithms are already our next DJs.

But not only are these algorithms able to curate music to our likenings; they are also able to flawlessly mix our favorite tracks together. Recently, a Spotify playlist was born that tests an automixing feature with the help of AI. The Drum And Bass Fix playlist seamlessly beatmatches two tracks when shuffle is switched on.

Not into drum and bass? Then try curating your own beat-matched set or mashup by using Rave DJ. This online application allows you to upload a YouTube or Spotify playlist with the use of algorithms. It then creates a smooth mix of even the most obscure track combinations.

Naturally, tech giant Google also engaged with algorithmic advances within the electronic music industry by developing an AI synth named NSynth. This open source synthesizer uses Google’s network to reproduce the qualities of sounds and instruments, which feeds its algorithms. Though based on neural networks, it actually comes as a hardware product with a touchscreen pad.

Will AI outmix humanity?

These tools may seem futuristic, but there are plenty of artists already utilizing AI to produce music. At this year’s Transmediale, UK DJ and producer Actress even granted his AI offspring complete artistic agency by giving it a stage name: Young Paint. Together, they enacted a live audiovisual performance that was mostly based on real-time improvisation, but they also captured some collaborative ventures on a mini-album via his new label Werk__Ltd.

According to electronic musician Olle Holmberg, it is just a matter of time before we will be following AI DJ’s and producers on social media, after attending our favorite algorithmically driven gigs – which is basically already happening with the advent of virtual influencers.

Based on the semantic traits that can be found in Hardwax’s database of DJ names, Holmberg recently published a list of DJ names generated by an AI. Though a DJ name might seem trivial, it does show that AI is capable of mimicking and further developing our club experience based on our current ideas of what clubbing should be like.

Team human

There is an uncanny objection to these kind of technological advances, assuming they would violate our authentic ‘humanness’, when in fact it is in our very human nature to be technological. Speaking, writing, reading counting, singing – these are all cultural technologies; so are DJing and producing.

The cycle that drove us from drum kits to drum machines is the same evolutionary force driving humans to interact with AI in creating new musical works of art. Within this framework, AI basically is our next nature’s cultural technology.

Scholar and electronic music composer Holly Herndon, who built an AI recording system to help with her latest album, addresses the pervasive narrative in which technology is dehumanizing and instead proposes to ‘run towards’ technology, but on her own human terms.

This brings us to the crucial debate revolving around AI: we often forget how algorithms are technologies developed by humans. If algorithms become dehumanizing vehicles, they can only be so because the human system made them that way. 

Enjoying this story? Show it to us!


Share your thoughts and join the technology debate!

Be the first to comment

More like this