GHOST WRITING
Why Does A.I. Music Pull The Wool Over Our Eyes?
THIS WEEK, a song by the country-pop group Breaking Rust reached No. 1 on Billboard’s list of top downloaded Country songs. The song “Walk My Walk” is, like a great number of songs Billboard rewards each week, pretty banal and predictable. The lyrics are a clićhed vision board of working class sentiments (“Been beat down, but I don’t stay low, Got mud on my jeans, still ready to go.”). The vocalist has a similarly gritty, beaten down quality to his voice. The song is also just vague enough that I could see it being used, easily, for both a Jeep commercial and an NFL promo. But I’m not entirely sure how to refer to the vocalist, since the singer does not exist. Breaking Rust, for that matter, is not a real band. Breaking Rust, and their (his?, it’s?) number one song, is an entirely A.I. generated entity and makes “Walk My Walk” the first song to successfully capture a top Billboard slot without a single human fingerprint present.
At this point in time, we are all more familiar and acquainted with A.I. generated visual art than with A.I. generated music. The proliferation of this uncanny valley A.I. art has seemingly taken over a great many social media feeds. Facebook, Instagram, Twitter, and TikTok have become cesspools of what has been labeled A.I. slop—generated images that have no real function or purpose other than to simply exist in an endless scroll. Open A.I., ever one to make a bad thing worse, launched a new app called Sora that just cuts to the chase and serves the user a trough of bottomless A.I. slop.
But A.I. generated music has proven to be far more difficult to detect than A.I. art. A.I. art, with its unnervingly smooth features, bright unnatural coloring, and, at times, extra appendages, is fairly easy to spot on a first glance. Not so much the case with music. A recent study by the polling firm Ipsos asked 9,000 people to listen to three clips of music, two generated by A.I. and one created by a human. The results found that a staggering ninety-seven percent of listeners could not tell the difference between an A.I. generated song and a human made one. People are listening, whether they know it or not, to more and more A.I. music. In January of this year, one in 10 tracks streamed each day were entirely AI-generated. Last month, that percentage skyrocketed to over one in three, nearly 40,000 per day. People are listening, unknowingly and unwillingly, to a fair amount of robot music.
Why is it then that so many of us can be fooled by bland, meaningless music and not by its visual equivalent? It could very well be that our ears, bluntly, are not as trained as our eyes. From an evolutionary perspective, we train our eyes during all waking hours of the day, to observe danger, falsities, and oddities. While driving this morning, I took a moment just to reflect on all the visual cues I had to process on my commute just to stay safe. Aurally, short of hearing a loud bang or a scream, we are rarely aware of our surroundings. The abstract nature of music, the intangibleness of it to an average listener, makes it hard to hear the musical equivalent of extra fingers. A slightly off chord, but one that is still rooted in the simple musical grammar of the song, is harder to pinpoint than an image of a cartoon child with unnervingly dead eyes.
We have also hung our ears up in recent years. Generated playlists curated by streaming services like Spotify and Apple Music have shifted our listening experience from an active one to a passive one…to now an almost unconscious one. A steady stream of music playing in the background—whose goal is oftentimes to help a listener study or relax or “chill”—has become aural wallpaper, the same kind of distant hum our refrigerator makes. This new function of music in our daily lives has opened up a vast avenue for A.I. generated material to seep through. If it really doesn’t matter what is playing, as long as something is playing, then it really doesn’t matter if that music was created by a human or not.
When the inevitable conversation arises with music students about this new musical landscape, they tell me that they would never listen to A.I. generated music. “It doesn’t have heart,” they often say in class. When I follow up with the question of what determines if a song has heart or not, there is a long silence. But in those same class discussions, they also admit to how listening has become an almost unconscious act for them. During one class discussion, a student confessed that the main reason he listens to music is that he is afraid to be alone with his own thoughts. When I asked who else agreed, the vast majority raised their hands. The ubiquitous use of headphones in public and the constant hum of playlists in dorm rooms do not reflect an obsession with music, but rather a desire for relief from introspection. Music, it would seem, has become mainly a source of distraction, made by both humans and machines—even to music students.
§
IN HIS 1934 BOOK OF ESSAYS Beyond the Mexique Bay, the British writer Aldous Huxley lamented that the collision of swift technological advancements and the quick and covenant production of art had, in turn, churned out an astonishing amount of meaningless garbage. “The proportion of trash in the total artistic output is greater now than at any other period….the population of Western Europe has a little more than doubled during the last century. But the amount of reading—and seeing—matter has increased, I should imagine, at least twenty and possibly fifty or even a hundred times.” A few months, Walter Benjamin argued the same in his essay “Art in the Age of Mechanical Reproduction.” That same year, Lewis Mumford made a similar point in his book Technics and Civilization. Just two years later, Charlie Chaplin turned all these technological anxieties into comic gold in his film Modern Times.
We now live, arguably, in a reflected time with the mid-1930s—a time of precarious economic uncertainty and cocky technological inevitability. Huxley, Benjamin, Mumford, and Chaplin were reacting to the the industrialization of nearly every aspect of modern life. Artificial Intelligence, like the mass producing technologies of the past, has bombarded all of us with a similarly high “proportion of trash,” now in the form of that meaningless noun “content.” Short of any regulatory action or social uprising, we all may be entering a kind of artistic industrial revolution. It won’t be that live performance or human created art will cease to exist, it will just become a kind of luxury good. The Industrial Revolution of the late nineteenth and early twentieth centuries took us from skilled woodworkers who made tables, chairs, and cupboards to IKEA and Wal-Mart. We may be facing a similar trajectory with music and art.