- The Analytical Songwriter
- Posts
- The Future of Music: Man or Machine
The Future of Music: Man or Machine
or: How I Stopped Worrying and Learned to Love the Robots
The subject of AI and the capabilities of AI chatbots has been hot in the news lately for a good reason. There isn't a single creative industry that doesn't stand to be affected by this emergent technology, and the potential effects are still unclear. Techno-optimists gaze enthusiastically into black mirrors, predicting the total automation of the world. At the same time, Neo-Luddites issue online screeds lamenting the loss of all humanity in the world. Amidst the bedlam and caterwauling of these online camps, a small group of people is skeptical of both claims. The concerns regarding the use of AI may be valid. Still, it is also true that we may assign greater capability to AI technology than is deserved.
The prognostications of online doomsayers have been no quieter in the discourse on music production and songwriting. In an interview with Rick Beato, Billy Corgan of Smashing Pumpkins remarks, "We're on the verge of AI systems that will literally make learning to play an instrument completely redundant."
Rick Beato responds, "Maybe..."
To Corgan's credit, he clarifies that while the prevalence of AI in music production will have a more significant impact on the state of pop music, it will never replace the personal journey conveyed by all musicians in their art. This observation cannot be overstated.
Beato provides a QED of Billy Corgan's argument in a follow-up video. He explains that the introduction of grid-based metronomic time and autotune paved the way for simulacra to dominate the music market by training the average music listener to expect perfection in records unattainable before the advent of the drum machine. Through this adjustment of collective taste and sensibility, a new aesthetic amenable to the contrivances of AI systems has emerged. These preconditions, coupled with the general disregard for lyrical content, Beato argues, make the prospect of AI-dominated pop music possible and likely.
Cooler heads like Benn Jordan Point out that the AI revolution still has a long way to go before it can even approach simulating the intricacies of human artifacts. In his video, AI Art: A Solution Without a Problem, Benn Jordan demonstrates the inconsistencies of generative outputs using mangled "stock" photos and poorly written script dialogue. Jordan attributes the cataclysmic conjectures of the music community to affluenza or the single-minded desire of an individual to gain wealth or status. He appeals to the fact that most critics of AI technology likely do not have careers in music and are unlikely to have said careers. Even if they did, Jordan argues, the industry does not value the labor of the average working musician anyway.
Jordan's most interesting point, by far, is his assertion that the concept of intellectual property is a construct designed to impose scarcity on something that, in reality, is relatively abundant. I believe the appeal to human innovation to be his strongest case for why AI cannot destroy music. However, the analysis is incomplete and not for lack of effort; Jordan glances at issues at the core of this problem but likely cannot articulate them in a 20-minute video. Appeals to innovation notwithstanding, a few questions arise when further considering the implications of Benn Jordan's video.
What is it that humans do that cannot be replicated by machines?
What is the role of art in the world anyway?
What are the incentives to replace human musicians?
These are questions too massive for a response in 2500 words. Still, I hope to grapple collectively with these questions and negotiate answers leading to a better understanding of art. In considering these questions, we may find that the threat of AI looms large only in a world that commodifies the arts and espouses entertainment as the supreme form of expression.
Oh, The Humanity!
The songwriting process is a strange thing. When you hear the greats talk about it, you realize it's equal parts drudgery and mystery. Colin Meloy had Chat GPT whip up a song in the style of his band, The Decemberists, and the results were interesting but ultimately underwhelming. The song had a coherent structure and chord progression and featured mostly intelligible lyrics. However, in the final analysis, Meloy arrives at the opinion that what chat GPT lacks is any amount of intuition.
The question arises: "What even is intuition?"
When we reach beyond the conscious rational processes of our mind and reframe problems creatively to arrive at solutions, we use intuition. But the mechanisms that drive this are still opaque to us, or else we'd likely have created machines capable of intuiting. The process is necessarily inefficient. And I assert it is necessarily so because humans already have an incredibly efficient way of communicating our inner lives through language and declarative statements. The reality is that what Tolstoy has to say in War and Peace can only be said in many words. Meloy can only gesture toward the chagrin of a downtrodden benchwarmer through its dramatization in The Sporting Life. I have a hard time believing that any artist would willingly subject themselves to the anguish of an artist's life to convey what they could state in a sentence.
Is intuition ultimately the struggle to divine meaning from seemingly random events? Is it the act of integrating our experiences into a cohesive narrative that we then term self? Is it the Christian God trying to bring us closer to the image he has of us in his infinite mind? We may never answer this question, but I have to believe that what answers we may find will not be reduced to only so much code.
What's Art Got To Do With It?
We must admit that the role of art in society is only slightly clearer than the fact of art's existence. Ask five people around you what art's purpose is, and the answer you'll likely get is some version of "art is self-expression." Don't get me wrong; this answer is compelling. It's also a pretty new answer, as humans are concerned. It was only 800 years ago that The Catholic Church governed the rules of musical composition. The idea that the composer would write music for anything other than God's glory was considered suspicious. Even peasants expected the troubadours who made music outside the church to sing about specific formulaic subjects.
We often view these constraints as moral injunctions on the part of oppressive power structures that seek to control the masses through media. While it's hard to dispute that this has been a factor since time immemorial, we are often blind to the fact that technology has stolen artistic merit from other areas of life. Mass production practices continue removing aesthetic considerations from furniture fabrication, cooking and dinnerware, and clothing. In the course of this streamlining and efficiency, we forget that these things were considered artifacts with aesthetic value at one time.
We derive the term art from the Latin ars, which refers to a practical skill or form of learning. The Greek cognate of the term, artyein, offers even more clarity when considering that it is a verb meaning to prepare or arrange. The shortcomings of tools of the past required artisans to develop compensatory skills before the prevalence of industrial efficiency. They construed this skill or learning as an art. In addition to producing artifacts, the artisanal craft contained stories of the culture that produced it by establishing itself as a tradition. By applying the term ars to music, we recognized the capability of song in society to perpetuate the stories of a our culture.
All this to say that art, and by extension music, existed to bring people into deeper participation with their culture. The only way to hear music was to be near it as musicians played or to play it for oneself. Beyond that, dances were a communal activity that allowed further participation of those who were not themselves musicians. The efficiency of the electronic replication of music has allowed this art to penetrate deeper aspects of our lives. In the modern era, access to any music of any time allows us to imbue the music with a personal sense of story----we are empowered to build a soundtrack for our lives in a way that our ancestors could never fathom.
The sound recording eliminated the need for a live musician. Doing so opened the door for legions of musicians to share their voices. By the 1960s, the album had become as culturally significant as the novel. As the internet became more ubiquitous, digital distribution showed that artists could speak directly to interested communities and find their audience. The proliferation of markets has led to an outpouring of music genres and sub-genres. While an AI can replicate the music of a given style, it cannot innovate in the same way that a human artist can. This fact puts AI at a complete disadvantage in today's music market, where listeners many listeners expect increasingly idiosyncratic options.
Efficiency and progress are ours once more
Industrial efficiency is an unseen force in the arts. Our industrialized model of media production is so entrenched that we tend to imagine that being an artist was always driven by some iteration of it. This presumption is evident when we speak of the artist as an occupation. The idea that there are swaths of people who strive to be artists to generate income would have been unimaginable to someone in the 13th century. Industrial efficiency has given scores of people the option to profit from what they love doing. The drawback is that there has always been someone who profits even more than the artist. Those who facilitate the money-making apparatus have every incentive to manufacture more music more quickly and efficiently. The industrial model applied to the arts is the dominant reason one could ever fear AI taking over music as we know it. Without it, there is simply no incentive to replace human musicians.
The hyper-optimization of music production is contingent upon the vested interests of corporations to produce more media content at the lowest possible price point. In a world where corporations dominate the entertainment industry, it is nearly unquestionable that AI would come to dominate music composition. If it were the case that this technology had arisen in the 70s, I think that the odds of an AI corporate music takeover would be greater. The rise of independent creators is a hopeful indicator that what many people want from music and other arts is a personal connection. In his video, Benn Jordan mentions that he makes more money from Patreon and Bandcamp than any other source and proposes that human empathy drives the new audience-driven profit model. While this may be true, audiences may also desire greater connection and participation in the media they consume. Music listeners may be instinctually harkening back to an era when music was participated in directly as a player or as a dancer. This type of craving for participation is something that AI cannot fulfill in any meaningful sense.
AI and generative music composition have been a looming boogeyman for quite some time. It is crucial to realize that the advancements in the field have not yet equated to a superior product to human capabilities. Even ChatGPT has flubbed on basic facts and mathematical problems. The reality is that there is a lot of hype surrounding these technologies, and we're still probably years away from legitimately worrying about the replacement of humans in artistic endeavors.
Despite the glaring issues with this technology, it is quite understandable that there is general anxiety regarding the automation of the arts and culture. We have already seen so much of our existence automated away in the march toward greater efficiency that it's hard to believe the day won't come when machines eliminate people from all "human" endeavors. For those who are concerned about the erasure of humans from music, I propose several simple solutions
Learn how to play an instrument.
The death of the amateur musician is something that David Byrne has lamented in his book How Music Works. The necessity of people making their own music endowed it with the spark of humanity. The ethos of rejecting corporate entertainment in favor of local and regional media has continued to sustain the DIY punk scene. This type of localization of the arts can single-handedly thwart the attempts of corporations to homogenize entertainment. No one gets rich doing it, but the ownership people feel as a community is its own reward.
2. Support local musicians and artists.
Prioritizing local musicians over international acts is another way to feel more connected to music. These artists share time and space with you, and, likely, their stories aren't dissimilar from your own. We don't have to reject artists from other areas, but we should recognize that international artists are necessarily abstract to us. This type of abstraction facilitates the cult of personality that gives way to celebrity culture. Because these people are untouchable, they cease to be people and become demigods beyond average human capacity. Seeing local artists in this light is impossible because you can get to know them as people.
3. Be skeptical of the belief that more efficiency is better.
The impulse to chase after efficiency is something that we all must negotiate. This impulse drives us to be hyper-productive and has knock-on effects on our mental and physical well-being. Allowing some spaces in life to be less efficient can be a form of protest. Refusing to buy the hype that the latest technology is able to deliver us from inefficiencies will allow you to speak to people in a way that reminds them that this technology comes at a cost. You may not slow the adoption rate of a technology, but you can help people to see that no technology is a panacea to the world's problems.
Over 200 years ago, the Luddites smashed the industrial machinery throughout England to protect the jobs and skills of textile workers and other tradesmen. Many musicians consider themselves at a similar inflection point without realizing the profound lack of scarcity in media markets. While this saturation makes earning a living as a musician more difficult, it has also increased opportunities for people to gain recognition as musicians. The future of media consumption seems to be moving away from a consumer model to an engagement model. In doing so, new doors will open for more human-centered modes of operation. The idea of a Borg-like assimilation of music may make for compelling science fiction, but it does not yet approach the realm of convincing reality.