This essay was written for a course on Digital Technologies and Power in November 2023.
As I sit down to write this blog post, Mac Miller’s soulful melodies wash over me,
painting the picture of a serene campfire with glowing embers in a snug cave amidst a
raging blizzard. His final record, Circles, spins on my turntable, a beautiful and
introspective yet haunting album that addresses the cyclical nature of patterns in his life
intertwined with a sense of catharsis having finally found his stride after years of
heartbreak and feeling lost. Across these almost 50 minutes of ascension, he takes me
on a journey, offering an intimate look into his life, allowing me to feel the weight of the
world on his shoulders in Good News to the feeling that everything that goes wrong is
somehow his fault in That’s On Me and finally the sense of stability he feels after finding
himself and living life on his own terms in Once A Day. This album, with its ability to
change my perspective of life, and countless albums like this, is why my favourite
human invention is music, an artform of limitless expression and emotion synthesised
out of a blank sonic canvas. Music has held the power of being able to pass down
stories and ideas that were never written, to inspire and evoke feelings, to channel
connections to gods and so much more.
Just as a liquid spreads out and fills its container, so too does music, except the
container in question is human culture and society. This is evident from the
good-vibes-only, carefree funk that dominated the 80’s to the more digital, darker,
anti-social trap found in the early 2020’s. Music is constantly evolving, imitating life as
life imitates it, having a say on politics, fashion and language. It is a uniquely human
creation, intended for humans, created by humans and about humans. However, a
recent concern in the music world is the usage of Artificial Intelligence to create new
music. This inclusion can range from using voice changers to sound more like a
particular artist to having entire melodies and lyrics written by AI. This has been a widely
criticised phenomenon, with some justifiably calling it the “death of artistic expression”.
The biggest criticism of using Artificial Intelligence to generate music is that the personal
aspect of music is lost when an algorithm that hasn’t lived through human experiences
tries to replicate what human artists put out, drawing only from what you feed it rather
than experimenting and creating novelty. An AI algorithm can never know what it is to
go from growing up on the streets of Compton, California, rife with abject poverty and
gang violence to living out dreams as Kendrick Lamar describes in Good Kid, M.A.A.D
City or the overwhelming, grand nihilism presented on Have A Nice Life’s magnum
opus, Deathconsciousness. Listeners to AI generated music will never feel the human
connection to the artist behind the music that they feel with music written by humans
because, well, there is no one behind the music, just numbers determining sequences
to please its reward mechanism.
There has also been large ethical concern over creating AI musical projects involving
the style and voice of dead artists. I often like to think that an artist has two lives, one
being their mortal, physical manifestation and the other being the legacy arising from
the body of work they produced over their lifetime. It is precisely this legacy that is being
tainted by AI generated music without any consent or even knowledge from the artists.
This is well out of scope for the vision that the artist held for their body of work and is
unwelcomed by most artists who are currently alive. To most critics of releasing AI
generated music, it seems that the record labels that make decisions on the
posthumous releases of artists are far more interested in profiting from the artists’
demise rather than caring about what the artists would’ve wanted and the vision for their
work, often releasing mediocre or outright terrible bodies of work that use AI in order to
have something to present to the public as a completed project. A legacy that could be
built up in decades could all come crumbling down in a matter of months as it gets
increasingly more difficult to tell what the artist made and wanted to put out for the
public to listen to.
This also gives rise to the problem of a sea of endless content, focusing far more on
quantity than quality. The power of being able to generate a song or an entire album
from a simple prompt means that people don’t get the opportunity to truly appreciate
and understand the meaning of the music, or lack thereof, in AI generated music before
moving on to entering their next prompt once the novelty of the music wears off. There
is a reason that artists such as Juice Wrld had only 4 studio albums released despite
reportedly producing thousands of songs over his lifetime; a lot of the songs were not
able to pass his high standards for being published and it doesn’t give the listeners time
to really understand and dissect the work he had already put out. We have already
started to see a transition from music as something that has inherent meaning and
should be savoured to something commodified, existing to make fortunes off of with
more singles with shorter run times and EP’s being put out rather than LP’s and singles
with longer run times that take the time to illustrate and detail their purpose.
This is not to say that AI has absolutely no place in music, but rather that music should
remain a human creation, used to make sense of our world as we see it, keeping the
culture in the hands of the people that are affected by it rather than have a non-human
entity dictate how we should feel, dress, act and think.