Being and Creation

The age of artificial intelligence promises many things, and abundance is one of them. When OpenAI’s DALL-E or Midjourney first produced entirely AI-generated photos and videos, we thought that, like the promise of abundant goods, we would also be promised an abundance of art and culture in the age of AI. This would lead to mass layoffs in the entertainment sector and whatnot, but imagine the works that would come from a 10x Shakespeare or a 10x Picasso (i.e., artists with ten times the aura and ability as a Shakespeare or Picasso)! We’d surely enter a new, 10x’d enlightenment that would also improve the human condition culturally and artistically! But with the production of AI slop, or nonsensical works of literature and art by AI, it seemed clear that the next great artist couldn’t be AI, and that there is some substantive difference between an artist who’s an AI model versus a human. More specifically, the production of slop is evidence that AIs, on their own, couldn’t (possibly ever) create works of art that’d be historically weighty or important, because it lacks two components: an ethical and willing faculty, the former of which is found in humans and the latter found in all living things. Without those two necessary components, no work of art can sustain the test of time because they are neither historically relevant nor executed willingly with a trace of life.  

We can understand the phenomenon of “slop” through the modern collapse of values, where products and parts of society float free of inherited value structures and lack significance. The Portuguese writer Fernando Pessoa, in his Book of Disquiet, diagnoses modernity as the collapse of shared frameworks that once made creation meaningful. In particular, he says that the slow withdrawal of a central, organizing value that was once embodied in the Christian God has all but collapsed into a single, esoteric framework that puts man’s interests first. For instance, each domain of human activity, like art, commerce, war, and politics, was thrown back on itself, forced to pursue its own internal logic with radical, self-referential thoroughness. Instead of doing things in the service of God or a supreme value, art became art for its sake; business became business for its sake; war became war for its sake. What goes wrong when each domain operates purely for its own sake is existential. A system pursuing its own logic to the absolute has no internal mechanism for restraint, because restraint would require answering to something outside itself. 

In the same spirit, when a model produces a poem or a painting, it operates within an autonomous system that is pattern completing to optimize for coherence as ends in themselves without paying any attention to whether the result serves any larger framework of meaning. The result is perfectly resonant with Pessoa’s “l’art pour l’art”, or an esoteric and technically polished work but unmoored from any value-positing center; in other words, nowhere in its process of creation did a subject choose this over that because something was at stake. Thus, AI art is not bad because it lacks polish or does not make sense. Rather, anything that it produces by itself is devoid of a central value system that checks against whether this word or that brush stroke is meaningful in the work as a whole. 

If slop is the product of a centerless system, then the first thing a centered one requires is an ethical faculty. To be sure, I am not talking about ethics in the narrow moral sense, but the capacity to discern among infinite possibilities the things that are worth doing. The philosophical groundwork for this claim lies in Heidegger's existential analytic in Being and Time. For Heidegger, the human being (Dasein) is that subject whose own being is an issue for it. Dasein is always already thrown into a reality and projecting itself toward future possibilities, and this structure, what Heidegger calls care (Sorge), is the ontological condition by which anything shows up as significant at all. Something is existentially near to Dasein when it matters, when it is relevant to Dasein's concerns and possibilities. And because Dasein's existence has the character of "mineness" (Jemeinigkeit), as my being is never a matter of indifference to me, I encounter a world in which some things call for attention and others recede into irrelevance. The ethical faculty, in the sense relevant to making historically weighty work, begins here, in the fact that a human creator stands in a world where things are already at stake for her, where choices carry the weight of a life that is finite, situated, and irreversibly her own.

Charles Taylor develops this Heideggerian insight into a more precise account of human agency. In his essay "What is Human Agency?" and later in Sources of the Self, Taylor argues that what is distinctively human is the capacity to evaluate our desires, what he calls "strong evaluation." To weakly evaluate is to calculate between alternatives on the basis of simple preference: I want this more than that. Strong evaluation is qualitatively different. It involves judging our desires against a background of distinctions between things recognized as being of higher or unconditional importance and things that lack this worth. The strong evaluator asks whether what she wants is worth wanting, what kind of life it expresses, what quality of existence it sustains. Taylor calls human beings "self-interpreting animals," a formulation indebted to Heidegger's analysis of Dasein as a being that understands itself through its possibilities, and insists that to lose the framework of strong evaluation would be to lose what is peculiarly human in agency altogether. It would be what Taylor describes as a pathological state, an identity crisis in the deepest sense: not knowing who one is because one can no longer distinguish between what matters and what does not. Without strong evaluation, the self has no orientation toward the good, and without orientation toward the good, the self cannot create anything that carries the weight of its own significance.

Discernment alone of what matters or what is important, however, leaves no mark on the world. One can perceive exactly what needs to be done and never do it. The second component is the willing faculty, the force that translates discernment into actuality through sustained effort, sacrifice, and resistance. Schopenhauer, in The World as Will and Representation, was the first to identify will as the metaphysical substrate of all existence. He describes it as a blind, unconscious, aimless striving that manifests in every phenomenon from gravity to human desire. The will expresses itself through ascending grades of complexity, from crude matter to plant and animal life and finally to human intellect and self-awareness, and Schopenhauer insists that "will" here is not merely human volition or conscious desire but a metaphysical principle underlying all of nature. In this sense, will belongs to all living things. What matters for our purposes is the phenomenological dimension of this striving: will as the lived experience of pushing against something that pushes back, like stones that resist the sculptor or language that resists the poet. To will, then, is to meet that resistance and persist, and in persisting, to leave a trace of one's life in the material. 

The ethical and the willing faculties are each necessary conditions for historically weighty work, but neither is sufficient on its own. To see why, consider what happens when one operates in the absence of the other. Ethics without will is Kant's famous "good will," pure intention that shines, as Kant himself put it, "like a jewel for its own sake." The person of impeccable moral discernment who never acts, never builds, never risks execution in the resistant material world, leaves behind nothing but the memory of fine principles. Kant acknowledged that the good will has its full worth in itself, independent of any outcome it produces, and there is something revealing in that formulation. A will that exists for its own sake is precisely one that has withdrawn from history, from the difficult and uncertain business of making things real. It leaves no trace in the world because it never entered the world. And in this way, ethics without will resembles the idealist who perceives exactly what ought to be painted or written or built, who possesses genuine orientation toward the good, and who never picks up the brush. Strong evaluation without the striving to actualize it remains purely internal, a private relationship between the self and its values that never becomes legible to anyone else.

Will without ethics is the condition Pessoa diagnosed in his "Disintegration of Values." Art for art’s sake, business for business's sake, war for war's sake: each is an exercise of tremendous will, tremendous energy and follow-through, directed by nothing beyond its own internal logic. The painter who pursues painting's logic to its absolute conclusion produces work that is, in Pessoa's account, completely esoteric and comprehensible only to those who produce it. There is no shortage of striving, effort, or even sacrifice in such work. What is missing is the ethical dimension, the strong evaluation that would subordinate all that energy to something that actually matters. The result is technically accomplished but historically weightless, because will without ethics is will without orientation, and without orientation, the most prodigious execution can only circle back on itself.

The question, then, is whether machines can ever possess both faculties simultaneously to produce meaningful, historical work. My answer to that question is that no, they cannot, due to purely structural reasons. First, AI cannot perform strong evaluation because it has no self whose identity is constituted by qualitative distinctions of worth. In Heideggerian terms, it has no “mineness”, no finitude, no thrown situation from which certain possibilities acquire urgency and importance. Instead, its “choices,” or probability distributions, are governed by statistical weight—the paradigm of weak evaluation. And so its products, however polished, are the products of a being for whom nothing matters, offered to a world in which nothing is at stake for it. Without such an ethical faculty, one can generate endlessly and never once produce something significant.

Secondly, AI cannot be said to face any resistance in its act of creation, as it does not experience a sense of lack that drives all willing and striving. But, one might object that a form of AI, such as a language model, faces resistance of its own kind. Namely, token limits constrain its output, reinforcement learning from human feedback shapes and restricts what it can say, and compute budgets impose scarcity. These are genuine constraints, and it would be dishonest to pretend otherwise. But constraint and resistance are different in kind, not merely in degree. A constraint is an external limitation imposed on a process that is indifferent to it. A chatbot running out of tokens is like a writer running out of ink. But the model does not experience its token limit the way the writer experiences the resistance of struggling to choose the right words or ideas in his work. In other words, AI’s token constraint provides no felt gap between what it means and what it has managed to say, because it does not mean anything at all. To make, then, is to struggle against the material toward some value that the maker has chosen and that the material does not yet embody. Since the will is striving that arises from lack, and lack is something that can only be felt by a being for whom the absence of the desired thing registers as a deficiency, a language model cannot be said to produce with resistance, even if it produces under constraint. And because resistance in the Schopenhauerian sense is constitutive of will, the absence of resistance means the absence of will, and the absence of will means the absence of any trace of life in an AI product.

There is a further risk that deserves attention, and it concerns not what AI produces on its own but what it does to the humans who rely on it. Even if we grant that AI is fundamentally a tool, over-reliance on that tool may erode the very faculties that make human products historically weighty. The ethical capacity, the ability to discern what is worth doing, atrophies when one stops exercising it. The willing faculty atrophies in a similar fashion. If the experience of creative work is increasingly frictionless, if the resistance of the material is absorbed by the machine so that the human never struggles against language or form or structure, then the human loses the capacity for the kind of striving that leaves a trace of life in the product. The danger is not only that AI produces slop directly, but that it produces a kind of human who can no longer produce anything else. Ultimately, the faculties of strong evaluation and creative will are capacities that develop through exercise and deteriorate through neglect.

The honest position, then, is not that AI can never participate in historically weighty work, but that it can do so only when the human's ethical choices and willed effort remain the governing structure of the product. When the human artist uses AI the way a sculptor uses a chisel, as an instrument subordinated to her vision, her judgment, and her sustained labor, the product can still carry historical weight because the value-positing center is intact. The moment the instrument begins to replace the center, the moment the human delegates not just execution but discernment and struggle, at which point the product loses its claim to significance. And determining where that threshold lies, where assistance becomes replacement, is perhaps the most urgent question for anyone who makes things in the age of AI.