Frankenstein, Pollock, and the Algorithmic Void: The Limits of AI-Generated Art
Embodiment, Context, and the Struggle for Meaning in Artistic Expression
Embodiment, Context, and the Struggle for Meaning in Artistic Expression
My most recent article on Medium took a critical look at the actual output of AI Music Engine Suno.ai through the lens of my extensive experience as a musician and music analyst. With 25 years of touring the world, playing on hundreds of records as a professional session drummer, and 17 years analyzing music for Pandora’s Music Genome Project, I bring a unique perspective to the current discourse on AI and art. It is from this vantage point that I attempt to examine the recent emergence of AI-generated text, art and music. Beyond the technical sophistication of these systems, I aim to explore the deeper questions they raise about creativity, authenticity, embodied experience in artistic expression, and their potential impact on society and the human condition.
I Asked Generative AI Music Platform Suno to Write a Podcast Theme…Things Got Weird
AI, Music, and Coherence: What Are We Trying To Accomplish?medium.com
The feedback I received, both publicly and privately, regarding my last piece was overwhelmingly positive. One paragraph in particular seemed to strike a chord with many readers and will serve as the foundation for our current exploration of AI, art, and society:
My body, my mind, my entire being — the 25 years of professional studio and touring work, the ~40 years of intensive listening to music — all scream at me: this is a (PRE)CHORUS! Only a disembodied entity that has never experienced the movement of a body to the rhythm of a beat could make a decision like this.
My hope with this paragraph is to underscore the central role of embodied experience, which includes context, temporality (duration) and the metabolization of life into an understanding of appreciating and creating art. It is from this vantage point that we will delve into the notions of embodiment, context, temporality, and information. By exploring these four dimensions, we can better understand the fundamental differences between human and AI-generated art and why this gap has profound implications for our freedom and the very essence of what it means to create and engage with art.
Have you heard of the story of Frankenstein? I chuckled a little bit writing that question. The story of Frankenstein is so thoroughly woven into the fabric of Western culture that I’m not sure it’s possible to escape its influence. In preparation for writing this article, I asked my 7-year-old if they knew the story, and they replied, “Of course, daddy,” with an eye roll. We have never read the book or watched a movie adaptation together, but at some point in their roughly 2,500 days on this planet, they had absorbed the basic premise of this iconic tale.
But how did this work of art come into existence in the first place? Does it matter how Frankenstein materialized?
In my opinion, the most important question is why Frankenstein even exists. We are incredibly fortunate to have a detailed, first-person account of the processes that resulted in the materialization of one of Western culture’s most enduring narratives.
Mary Shelley, writing in the introduction to the 1831 edition of the novel, which was first published in 1818, states:
[…] As a child I scribbled; and my favourite pastime, during the hours given me for recreation, was to “write stories.” Still I had a dearer pleasure than this, which was the formation of castles in the air — the indulging in waking dreams — the following up trains of thought, which had for their subject the formation of a succession of imaginary incidents. […] but my dreams were all my own; I accounted for them to nobody; they were my refuge when annoyed — my dearest pleasure when free.
I lived principally in the country as a girl, and passed a considerable time in Scotland. I made occasional visits to the more picturesque parts; but my habitual residence was on the blank and dreary northern shores of the Tay, near Dundee. Blank and dreary on retrospection I call them; they were not so to me then. […] It was beneath the trees of the grounds belonging to our house, or on the bleak sides of the woodless mountains near, that my true compositions, the airy flights of my imagination, were born and fostered.
This is remarkable. What we see here are the seeds of a passion being sown and nurtured, in place, in time, and out of time, in the fancy of spectacular imaginations born of the affectual effects of the surrounding environment.
But this alone did not produce Frankenstein, the novel considered the first science fiction novel, the first modern horror story, and a pioneering work of Gothic literature. There are embodied experiences through time which must emerge:
In the summer of 1816, we visited Switzerland, and became the neighbours of Lord Byron. At first we spent our pleasant hours on the lake, or wandering on its shores; and Lord Byron, who was writing the third canto of Childe Harold […] But it proved a wet, ungenial summer, and incessant rain often confined us for days to the house. Some volumes of ghost stories, translated from the German into French, fell into our hands. […]
“We will each write a ghost story,” said Lord Byron; [….]
I busied myself to think of a story, […] I thought and pondered — vainly. I felt that blank incapability of invention which is the greatest misery of authorship, when dull Nothing replies to our anxious invocations. Have you thought of a story? I was asked each morning, and each morning I was forced to reply with a mortifying negative.
But then something happened, that could only happen in this very time and place:
Many and long were the conversations between Lord Byron and Shelley, to which I was a devout but nearly silent listener. During one of these, various philosophical doctrines were discussed, and among others the nature of the principle of life, and whether there was any probability of its ever being discovered and communicated. They talked of the experiments of Dr. Darwin,[…] who preserved a piece of vermicelli in a glass case, till by some extraordinary means it began to move with voluntary motion. Not thus, after all, would life be given. Perhaps a corpse would be re-animated; galvanism had given token of such things: perhaps the component parts of a creature might be manufactured, brought together, and endued with vital warmth.
Night waned upon this talk; and even the witching hour had gone by, before we retired to rest. When I placed my head on my pillow, I did not sleep, nor could I be said to think. My imagination, unbidden, possessed and guided me, gifting the successive images that arose in my mind with a vividness far beyond the usual bounds of reverie. I saw — with shut eyes, but acute mental vision — I saw the pale student of unhallowed arts kneeling beside the thing he had put together.
An unseasonably rainy summer, being stuck indoors, happening to be hanging out with Lord Byron, a congenial prompt to ‘write a ghost story,’ a late night talk about Dr. Darwin — the grandfather of the man who would become the famous Charles Darwin — spurred those flights of fancy in Mary Shelley, as she lay in her bed that night.
From this, the universe was gifted the story of Frankenstein, which has left an indelible mark on our culture. But could a generative AI machine produce such a work, with a similar profound impact on our culture, without moving through time, context, and subject to contingency?
I contend, absolutely not.
Currently, AI is trained on decontextualized and detemporalized data. These AI models, whether they generate text or music, use sophisticated neural network techniques, such as transformer architectures, to process and learn from vast amounts of data. For text-based models, the training process involves feeding the model a sequence of tokens (words or subwords) and asking it to predict the next token in the sequence. Similarly, music-generating AI models learn from large datasets of musical compositions, analyzing patterns and structures to create new pieces. By iteratively adjusting the model’s parameters to minimize the difference between its predictions and the actual next token or musical element, the model learns patterns and relationships within the language or music.
However, this training data is sourced from the internet, books, musical databases, and other static sources that lack the web of non-propositional schematic structures emerging from the bodily experiences we humans bring to every interaction. The model learns to recognize patterns and generate responses based on statistical regularities in the data, but this can never accurately represent our experience of the world for two reasons: (1) meaning in natural language begins with figurative, multivalent patterns that cannot be reduced to a set of literal concepts and propositions; and (2) these patterns and their connections are embodied and cannot be reduced to a set of literal concepts and propositions. As Mark Johnson argues in his 1987 book “The Body in the Mind,” meaning is derived from nonliteral (figurative) cognitive structures that are irreducibly tied up with the conceptual or propositional content attended to in these LLM models.
These issues of the internal representations used by these models, such as embeddings and attention mechanisms, are abstract and do not directly correspond to the embodied, situated nature of human cognition and creativity. While these techniques allow the model to capture complex linguistic or musical relationships, they do not inherently provide the model with a grounded understanding of the world or the ability to reason about physical objects, their interactions over time, or the emotional and sensory experiences that fuel human artistry.
As a result, while LLMs like ChatGPT, Claude, and Gemini, or music-generating AI like Suno.ai, can generate fluent and seemingly coherent responses or musical compositions, they lack the deep, context-dependent understanding that arises from embodied experience in the physical world. The responses and creations are based on abstracted patterns in the training data but lack the capacity for a true temporal reasoning within a contextual adapted environment informed by the life experiences that are central to human cognition and creativity.
But isn’t everything just patterns. What’s so special about the difference between say a Jackson Pollock output, and a Midjourny output that mimics a splatter painting?
Everything actually.
In the December 2002 issue of Scientific America, there is a fascinating and extremely relevant article to our discussion. In 1994, Richard Taylor, a physicist and painter, decided to put his career as a physicist on hold, and give his dream of being a painting a full-go of it. He went back to college and during a week long field trip in the North of England was assigned the task of going outside and painining what they saw. As chance would have it, a rare week long snowstorm made this task ‘impossible’ and Taylor along with some new friends schemed to assemble a large structure on a tree branch blown down from the storm:
One part of the structure acted like a giant sail, catching the motions of the wind swirling around it. This motion was then transferred to another part of the structure that held paint containers, and these dripped a pattern corresponding to the wind’s trajectory onto a canvas on the ground. (Taylor, 2002, p. 117)
The result? The storm left behind a Pollock! Taylor explains the insight he had from this:
Suddenly, the secrets of Jackson Pollock seemed to fall into place for me: he must have adopted nature’s rhythms when he painted. At this point, I realized I would have to head back into science to determine whether I could identify tangible traces of those rhythms in his artwork. (Taylor, 2002, p. 118)
Taylor’s serendipitous discovery, coupled with his background as a physicist working with fractals, led him to investigate whether fractals were present in Pollock’s work. During Pollock’s era, when he was actively painting, nature was viewed as ‘disordered’ from a scientific perspective, with no overarching governing logic to ‘natural systems’ like weather. It wasn’t until the 1960s that scientists began to study these processes in depth, ultimately discovering an underlying order within them.
In the 1970s, a new form of geometry was developed to articulate the patterns these ‘chaotic’ processes left behind. This new geometry was named fractals by their discoverer, Benoit Mandelbrot. Fractals are composed of patterns that recur on finer and finer magnifications, building up shapes of stunning complexity and beauty.
Taylor decided to study Pollock’s paintings using state-of-the-art computer analysis and was able to determine that Pollock’s paintings did indeed contain fractals. Moreover, when his paintings were studied within the temporal horizon of his development, the density of these fractals increased as Pollock refined his technique over time.
Pollock’s paintings were not merely the result of random splatters but were deeply connected to the artist’s embodied experience and his interactions with the natural world around him.
It is important to note that Pollock developed this technique after leaving New York City and moving into a small house on the East End of Long Island. In this natural setting, something was sparked inside Pollock. Like the spark that haunted Mary Shelley while she lay awake with her eyes wide-shut, Pollock’s body began to metabolize this new embodied environment of Springs, NY, initiating a new artistic vision.
In October 2023, I visited the Pollock-Krasner House and Study Center. The tour guide took us to the very first room Pollock painted in when they moved to Springs, NY, in 1945. As I stood there looking out the window, I was awestruck by the scene. This was what Pollock was taking in, and it eventually spurred him to yearn for more space to allow what was building inside him, a portal out. This led him to move into the abandoned fisherman’s shack on the property.
The lights, the air, the view, the spaces — these natural and embodied affective responses led to the fractals in Pollock’s work. This is significant as it shows that his paintings are not merely aesthetically appealing but also have a connection to the patterns found in nature. As Taylor points out, “Twenty-five years before their discovery in nature, Pollock was painting fractals” (Taylor, 2002, p. 118).
When I was in college getting my degree in Studio Music and Jazz Performance at the University of Miami, I was practicing 8 hours a day on top of my 21-credit course load. I slept less than 4 hours a day, and every single waking moment was focused on refining my technique.
One day, I don’t know what happened — well, I kind of do. I was beginning to have these thoughts of ‘why am I even doing this?’ ‘What is the point of learning how to play a groove in 15/16 time while playing a modified Afro-Cuban groove for the drumset?!’ I was walking across campus for my one mandatory non-music class that semester when I decided, instead of walking past the library as I usually did, to turn right and go through the doors.
I walked to the front desk and asked where the collection of art books was located, then strode over to the elevator. I stepped out on the 5th floor and moved around the stacks, randomly pulling oversized art books out and just looking at them.
At some point, I pulled out a very large book on modern art, and there was a section focused solely on Jackson Pollock. I had seen Pollock’s paintings at MOMA and was familiar with his work, but he was in no way a central figure in my life story. However, on this day, as I was flipping through the thick pages, I came across a series of his paintings from the ‘pre-splatter’ period, featuring scenes from Wyoming to New York. One of them depicted a seashore and a lighthouse, painted in perspective — a scene that any technically proficient painter could produce. It was at that moment that I had an epiphany: Pollock didn’t just wake up one day and start throwing paint on unrolled canvas in a shed on Long Island. He had immersed himself in the life of an artist, studying techniques, styles, and mediums. Most importantly, he had stood in these places and painted these scenes. He had moved through those scenes and metabolized them.
This realization shook me out of my funk, and I went back across campus to my next music class with a newfound vigor. I leaned into my practicing, but not for the sake of technique alone. No, I was toiling away for the ability to express effortlessly and authentically what I metabolized in the scenes I moved through.
At the beginning of my last piece, which I linked to at the top of this essay, I critiqued a theme song produced by Suno.ai. In that article, I included a quote from their “About Us” page:
Suno is building a future where anyone can make great music. Whether you’re a shower singer or a charting artist, we break barriers between you and the song you dream of making. No instrument needed, just imagination. From your mind to music.
After sharing this quote, I mentioned, “I am inclined to write a 3,000-word think piece just on that, but I’ll spare you for now…”
Well, here we are, 3,100-odd words later, and it’s time to tie everything together.
“Suno is building a future where anyone can make great music.”
No you’re not.
What Suno.ai and all the current instantiations of AI platforms do is cheat you, and humanity, out of the very essence of what it means to be a thinking, experiencing, tragic human being.
You see, Frankenstein, Number One, 1950 (Lavender Mist), Machine Gun — these totems to the human condition— can only ever materialize and have meaning in the context of the web of human social relations. Relations that exist in time, in place, in history, and informed by our embodied desires of where we want to go.
AI output, the ‘information’ it generates, has no inherent value until we ascribe value to it. It takes decontextualized, de-temporalized, disembodied datums of information, transforms them using multi-layered neural networks and optimization algorithms, and spits out an endless stream of noise. This noise requires the embodied, temporalized, contingently contextually sentient human being to judge whether or not what is being fed to them is valuable.
The relentless velocity at which these AI machines function subjects us to a deluge of output, compressing our lived reality into an increasingly constricted space. Each fleeting chance for insight, each moment ripe for profound discovery, is mercilessly obliterated under the sheer weight of AI-generated ‘content.’ Our desires, the spaces we need to breathe and contemplate — ‘the formation of castles in the air — the indulging in waking dreams…’ — are rendered obsolete, smothered by the unyielding demands of these machines.
“[W]e break barriers between you and the song you dream of making.”
Life is not about the song. Life is about the barriers we overcome in the creation of ourselves, and how we relate that journey to each other.
Mary Shelley, 17 years old, had her first daughter born prematurely; determined to prove the attending physician’s diagnosis of certain death wrong, she managed to nurse her to health, only to have her not wake from a nap 2 months later.
Mary wrote in a journal entry of a dream she had that would profoundly impact her ideas of rejuvenating life in Frankenstein:
“Dream that my little baby came to life again; that it had only been cold, and that we rubbed it before the fire, and it lived. Awake and find no baby. I think about the little thing all day. Not in good spirits.” (19 March 1815).”
Jackson Pollock, who started drinking at age 15, ended up in Springs, NY because of his drinking problem. He found peace, solace, and inspiration in the clear air and picturesque scenery of Eastern Long Island and got sober. Finally — after decades of barriers and failures — he surmounted them, resulting in the production of pieces of art so astounding that they transformed the world. Yet, at the height of his career, Pollock returned to drinking, and within a short time, he wrapped his car, his body, and the body of his mistress around a tree.
Art we care about manifests through humans living in tragic contexts, in tragic times, and in our tragic bodies. It is through this very struggle that beauty and meaning find their resonance and humans find their purpose.