Photo by @simbrock

Artificial Intelligence and Music

The Musicians of the 21st Century Plays On

My good friend and composer made me aware that the field of artificial intelligence was used to help finish of Schubert’s Unfinished Symphony.

Composers and musicologists have puzzled over this for nearly two centuries now, but it took Los Angeles-based film and television composer Lucas Cantor, with the help of some iPhone-based AI from Huawei. The cloud-based machine learning system analyzed Schubert’s composition style from some 2,000 pieces and made suggestions that would fit the unfinished piece based on what it learned.
–Writes Bobby Owinski in the blog Music 3.0

Earlier this year Warner Music became the first record label to sign a deal with an algorithm. The mood music app Endel, creates bespoke music soundscapes for users, and it is expected to create 20 albums this year.

However who could be better than Spotify to get this trend going? Already in 2017 they hired Francois Panchet a French professor and artificial intelligence researchers focused on teaching computers to make their own music. It was decided that he should head up Spotify’s new Creator Technology Research Lab in Paris.

The results don’t quite stand alone as anything you’d want to listen to on your commute, but like Pachet’s work, both companies say their efforts are less about replacing artists and more about supplementing the creative process with automation.
-Francois Pachet

Spotify acquired Echo Nest in 2014. The Echo Nest is a music intelligence and data platform for developers and media companies. This brought some of the brightest minds in acoustic analysis and machine learning to the company. They have built music curation features such as: Release Radar, Discover Weekly, and Daily Mix.

While Spotify has invested in AI for purposes of music discovery and distribution, this is the first time the company has turned to machines for help with the creation side of the music equation. So it begs the question if the field of AI meshes with music, can there be distinctions?

As such we could perhaps differentiate somewhat cryptically to AI-assisted music and AI-created music. Perhaps still we simply could call both music, as the way to express music has been changing throughout time? Yet it could be fun to differentiate between the two.

If you initiated an algorithm and it created a song or piece of music then it could be AI-created, and if you adjusted the music or changed it then it could be AI-assisted. However this begs the question of how much was done to change the song, and here we go into the details.

This is a problem with the notion of autonomy. If you wrote the algorithm or decided to apply the algorithm to a specific project is it then AI-created? As with any tool that is used in the production of music to produce sound can we say that it was guitar-created? You have to pluck the strings.

DJs who make remixes have been criticised for the fact that they are taking other people’s music rather than producing their own. However they are warping existing sounds through re-combining and re-purposing the music into a new piece of music. Can we critique these any more or less than those who have a limited set of chords in various combinations that seem similar to other songs?

Original: present or existing from the beginning; first or earliest. Apparently the first reference is in original sin “innate depravity of man’s nature,” supposed to be inherited from Adam in consequence of the Fall. As such if we want to be overtly religious about music and what it is supposed to be that creates a lot of questions of originality — often in connection with copyright. Rights management in music is its own field dealing with this topic.

Now, hits are written and produced on iPhones (Steve Lacey, The Internet). New technologies have always given rise to new ways of making music. The introduction of DAWs (digital audio workstations) in the 80s let artists step out of the studio and create songs anywhere.

The important question as music is changing is how to reward musicians for the work they have been doing. This is uncertain in these times as it has been before, yet requires people to work on behalf of musicians to make sure they are awarded if they do well. If an algorithm initiated by a computer scientist can sign on to create 20 albums it poses some intriguing questions.

This is day 20 of #500daysofAI, I hope you enjoyed it.

What is #500daysofAI?

I am challenging myself to write and think about the topic of artificial intelligence for the next 500 days with the #500daysofAI. It is a challenge I invented to keep myself thinking of this topic and share my thoughts.

This is inspired by the film 500 Days of Summer where the main character tries to figure out where a love affair went sour, and in doing so, rediscovers his true passions in life.

AI Policy and Ethics at www.nora.ai. Student at University of Copenhagen MSc in Social Data Science. All views are my own. twitter.com/AlexMoltzau