The Death of Artistry

?

I’m an artist. Specifically, I am a musician, and have been playing the guitar and composing across multiple genres for as long as I can remember.

Generally, throughout my musical career composing has always been a collaboration of artists. I would sit in a band room with my band mates tossing around ideas and slowly but surely shaping them into songs with lyrics, instrumental breaks and guitar solos, keyboard parts and more. Even when working alone I collaborated with a singer who helped ultimately craft the songs that went into the recording studio and from there onto CD and into the music cloud.

I couldn’t tell you exactly how I created a song. They always seem to just fall out of my mind kind of like leaves falling off a tree in the fall. A few years back I wrote a song called “Across the Sea”, and it wasn’t until much later that a friend of mine read the lyrics and asked me if I had intentionally written an autobiographical piece of music. I had not intended that to be the case, nor did I realized it as I was writing.

Creating music took skill. It takes skill. Not only does a musician have to spend time and effort learning their own instrument, but the ability to craft a song of any length is yet another developed skill that takes years to hone and sharpen. You might look at today’s artists as having come out of nowhere, but unless they are singing music that is written via the “Swedish hit factory” of behind-the-scenes music fame they spent quite a long time honing their own craft.

I am now ever more convinced that AI is killing the art of making music.

I was watching a YouTube video by Rick Beatto, and in this video he was talking about the creation of AI music via a website called Udio.com . Naturally I was curious, so I headed over to that website to generate some music of my own.

What this system provided to me was both shocking and amazing. It was soul crushing and yet filled me with wonder. It simultaneously depressed me and filled me with happiness. The artist in me suffered all the negative emotions; the music that it generated was every bit as good as anything that I could myself create or could hear online from a favorite artist. The consumer in me loved it, as it opens the doors for the creation of all kinds of new music from people who are not musicians.

Here is a link to my personal udio.com web page:

https://www.udio.com/songs/1kSxhn5RArwRnJbCKdi5Ee ?

Have a listen and be honest in your own evaluation of this music. Could you tell the difference between this, filled as it is with vocal harmonies and musical runs and changes, and a song that had been written by a band or a person?

No one wants to go to a concert hall to see a laptop performing.

We all go to see live music to hear it and at least in many people’s cases to experience a performance from artists who are playing instruments and singing into their microphones. The line about seeing a laptop in concert is obviously both humorous and tongue in cheek, but it does reflect a certain bit of reality. If I go to see Nightwish, or Epica, or any other band in any genre of music, I go because I am looking forward to hearing and seeing old and new music that they have written.

How much does my opinion change if what I have in front of me amounts to a cover band?

Instead of spending months and months writing music, polishing it in the studio, and delivering it to us from our favorite music service, what if notable bands like the ones above or your favorite bands simply become experts at prompting an artificial intelligence system into writing music that sounds like they do? When the artificial intelligence delivers a song in the style of “Band X”, the band that performs it on stage is essentially a cover band performing music that was written by someone else. Would we view Eddie Van Halen differently if songs such as jump, or hot for teacher, had been written by an artificial intelligence system and he simply performed them on stage?

I suspect a generational gap in how people view this conversation. Over time music has become more and more computerized even in its own creation; auto tune and the ability to use your computer to do what’s called “comping” across tracks means that everything can be inch perfect and pitch accurate. By the time you get to see an artist on stage quite a lot of what you might see is flown in via computer with the artist simply performing a single piece on their own or in some cases doing little more than dancing and moving while lip syncing. This is not what you see with some of the older acts (Motley Crue excepted), whose members play instruments live on stage and sing without live pitch correction.

Even as recently as five or so years ago when I was in the studio with Legacy, we recorded music by doing what was called a “scratch track” where all the band members play simultaneously their own parts for a particular song. Each individual band member might then go back and overdub their own parts, replacing something completely or simply adding in places that need correction. Guitar solos were laid in as overdub tracks, as were vocals and vocal harmonies. With a computer as your guide you can easily write a song and then record it across several hundred different tracks, mixing them all together and delivering a finished product.

This is an expensive and time-consuming process in the best of scenarios with band members unable to perform their parts being replaced by studio musicians who can come in and quickly deliver what is needed for a given song.

Why bother with any of that when you can have a song written, performed, mixed, mastered, and then delivered by an artificial intelligence-based system inside of 10 minutes? Why write checks across dozens of different people when you can pay $19.99 a month to get the same result?

The next time you step into an elevator or walk into a coffee shop or hear music piped in that’s supposed to be relaxing like that which you might hear in a grocery store, remember that all that music was at one point created by a person. Taxi.com is but one example of a music licensing site that exists solely to connect music creators with music purchasers. If the music creators are replaced with artificial intelligence, then all those music creators are out of a job and have no way to earn income from their music.

For each creator that gives up an artistic pursuit, you can bet that there are several who never bother to try in the first place because you can’t make a living wage competing against the artificial intelligence that can write and deliver a completed piece of music in a fraction of the time you would take and an even smaller fraction of the cost.

Part of the reason that music has always for me been such a wonderful and diverse escape is because the music is so diverse. Rock and roll used to describe everything that was rock and roll and now is littered with subgenres that themselves have subgenres and I think it’s fantastic! It seems that artificial intelligence as a music design tool can only lead to the homogenization of music, and if everything starts to sound the same then the artistry dies. When you listen to top 40 music there’s a reason that it all sounds the same; it all sells that way and it all appeals to the purchasing audience. There’s certainly a market for that and even an appetite for that, but does anyone want to wind up living in a world where music, generated by computers, all sounds the same?

We can’t stop the march of progress. We should not try and stop the merch of progress. However, we should also try to maintain support of the organic; we should see live music and support bands that we love by buying t-shirts and coffee mugs and jackets. There’s nothing wrong with being musically elite. There is nothing wrong with drawing a hard line between those people who are musicians and those people who simply create music by telling a computer what to do. Perhaps it is a bit of a purest sentiment but if you tell a computer what kind of music you want to listen to you’re not a musician, and you have no right to claim that you wrote those songs. You didn’t, the computer did, and the only reason it was able to is because of the work of millions of musicians across hundreds of years that has been used to teach those computers what it knows.

It saddens me to think that somewhere out there there’s a youngster who is saying to him or herself why bother to become a musician? I don’t need a guitar or a bass or a set of drums, I just need a powerful computer and a few bucks. I can be a famous musician as long as I know how to prompt AI. If we allow that to happen across the musical spectrum, if we allow that to replace music then Rick Beatto is right and then in the Year 2100 will still have radio stations playing Stairway to Heaven.