Concerning the Spiritual in AI and Art

Concerning the Spiritual in AI and Art

Jer Nelsen
South Carolina School of the Arts, Anderson University

Abstract

This paper examines whether generative artificial intelligence (AI or GenAI) can produce authentic art by engaging the theological and phenomenological accounts of artistic creation. Drawing on Wassily Kandinsky’s concept of innere Notwendigkeit (inner need), Madeline L’engle’s incarnational theory of art, and Makoto Fujimura’s notion of “culture care,” I argue that authentic art requires something AI structurally lacks: a consciousness reaching asymptotically toward meaning. AI systems generate outputs by calculating statistically probable arrangements of tokens; they possess no inner need, no lived experience, no spirit to incarnate in matter. I introduce the concepts of “participatory activation” (what viewers bring to any image) and “deposited presence” (what only ensouled makers can embed in their work) to clarify this distinction. Extending Rem Koolhaas’ architectural critique of “junkspace,” I suggest that AI-generated content tends systematically toward adequate emptiness, risking a cultural regression to the mean that may erode aesthetic discernment. I engage a counterargument, the computational theory of mind that Dennet and Hofstadter advance, while noting that philosophers have not yet resolved the “hard problem” of consciousness. Rather than offering definitive answers, I invite readers to dwell in the tensions between competing accounts of consciousness and creativity, suggesting a posture of deliberate slowness, patient attention, and embodied presence.

1. The Asymptotic Fidelity of Language

There is a peculiar frustration known to anyone who has ever tried to describe something that matters to them. The words come, but they do not quite fit. We reach for language and find it slipping through our fingers, approximating but never arriving, gesturing toward meaning without delivering it. This is not a failure of vocabulary or education. It is a structural condition of what language is and how it works. Words are never the thing itself.

The phrase "asymptotic fidelity" captures this condition with precision.¹ In mathematics, an asymptote is a line that a curve approaches but never touches, drawing infinitely closer yet never arriving. Language operates this way with meaning. Our words approach what we wish to say; they draw nearer and nearer; but a gap remains, unbridgeable by any accumulation of syllables. The poet knows this. The mystic knows this. The grieving person who cannot make anyone understand what they have lost knows this with terrible intimacy.

We can be relatively precise about quantities and physical descriptions. "The table is three feet wide" communicates with reasonable fidelity. But "I love you" or "I believe in God" or "this artwork moves me" strain against the limits of what words can carry. We say them anyway, trusting that our circling will suggest the center we cannot name, that the shape of our failure will somehow convey what success could not.

The visual arts emerge from this gap. What words cannot deliver, perhaps images can show. What resists the sequential logic of syntax might yield to the simultaneous presence of form and color, experienced all at once rather than unfolding through time. The painter, the sculptor, the photographer: each works in a medium that operates differently than language, reaching toward meanings that verbal expression cannot quite touch.

Wassily Kandinsky, the Russian painter and theorist who helped pioneer abstract art in the early twentieth century, understood this with unusual clarity.² Writing in 1911, he argued that colors and forms possess what he called an "inner sound"; a capacity to produce "a corresponding vibration in the human soul" that bypasses rational cognition entirely.³ The painter speaks where the poet falls silent, not because visual language is superior but because it operates by different means, reaching places that words cannot follow. Yet even here the asymptote persists. The image, too, circles its meaning without exhausting it. A painting of the Annunciation does not deliver the experience of divine interruption. It offers a surface, an occasion, a catalyst. The meaning kindles in the prepared viewer or it does not kindle at all. The artwork is not meaning itself but the possibility of meaning, requiring participation from a consciousness capable of response.

This is the first thing we must understand before we can think clearly about generative artificial intelligence and art. Human expression, in any medium, operates within this asymptotic condition. We reach toward meaning; we do not grasp it. We offer signs and symbols that might resonate in other minds, but the resonance depends on what those other minds bring to the encounter. Communication is always incomplete, always approximate, always a collaboration between the one who speaks and the one who hears.

The question that haunts our present moment is this: What happens when the one who "speaks" has no meaning to reach toward? What happens when the signs and symbols are generated not by a consciousness striving to express something but by a system calculating the most statistically probable arrangement of tokens?

2. The Inner Need: Kandinsky's Claim and Its Limits

Kandinsky gave a name to the generative principle of authentic art: innere Notwendigkeit, usually translated as "inner need." This is not aesthetic preference or stylistic fashion. It is something closer to compulsion. The artist creates because something within demands expression, and the work that emerges bears the mark of that demand like a fingerprint, like a signature written in the texture of every brushstroke.

"The artist must have something to say," Kandinsky wrote, "for mastery over form is not his goal but rather the adapting of form to its inner meaning."⁴ Technique is necessary but not sufficient. Virtuosity without inner need produces at best decoration: pleasant arrangements that catch the eye but leave the soul untouched. At worst, it produces emptiness wearing the mask of significance, form without content, the shell of art without its animating life.

Kandinsky developed an elaborate system for understanding how colors and forms communicate spiritually, each carrying what he called an "inner sound" that bypasses cognition entirely.⁵ His famous image of the "spiritual triangle" depicts humanity arranged in ascending segments, with artists standing at the narrow apex, perceiving truths that others will only grasp generations hence. Some of his specific claims have not aged well, particularly the mystical vocabulary he borrowed from Theosophy. But the core insight survives the dated trappings. Art that matters comes from somewhere. It is not merely calculated to produce effects but arises from genuine necessity, from something the artist must say because they cannot not say it.

We recognize this quality when we encounter it, even if we cannot precisely define it. Standing before a Rembrandt self-portrait, we sense the presence of a consciousness grappling with time and mortality. The painting is not merely technically accomplished; it is inhabited. Something looks back at us from the canvas, and that something is not reducible to the arrangement of pigments.

Makoto Fujimura, the contemporary Japanese-American artist and writer, extends this tradition in explicitly Christian terms.⁶ Fujimura speaks of art as "culture care," a practice of tending and nourishing the soil in which human communities grow.⁷ His own work, created through the slow techniques of traditional Japanese Nihonga painting using mineral pigments and handmade paper, embodies a kind of making that resists efficiency and speed. The labor itself becomes part of the work's meaning, a form of prayer that deposits itself in layers of gold and azurite. The artwork becomes what he calls a "liturgy of making," a ritual act that participates in the ongoing creativity of God.⁸

Not everyone will accept this theological framing. The philosopher George Steiner argued that all serious art operates under what he called a "wager on transcendence" – an implicit bet that there is more to reality than matter in motion, that meaning is discovered rather than merely invented.⁹ One need not be religious to recognize that the greatest artworks carry a quality that cannot be accounted for by analyzing their formal properties alone. Something exceeds the sum of the parts.

The question is what that something is and whether it can be present in works generated without human consciousness, without inner need, without the lived experience of reaching toward meaning and falling short and reaching again.

3. Incarnational Making: L'Engle and the Breath of Life

Madeleine L'Engle, best known for her novel A Wrinkle in Time but also a serious theologian of art, offered language for this question that cuts to its heart.¹⁰ "True art is incarnational," she wrote. "It carries the breath of life."¹¹ The theological weight of that claim deserves careful attention.

In Christian theology, incarnation refers to the central mystery of the faith: God becoming human in the person of Jesus Christ. The infinite takes on finite form; the invisible becomes visible; spirit assumes flesh. This is not merely a doctrine to be believed but a pattern to be recognized, a pattern that, for L'Engle, extends throughout creation and is especially concentrated in the act of artistic making.

When an artist creates, something analogous to incarnation occurs. An inner reality, a vision or emotion or intuition of truth, takes on material form in paint or stone or sound or word. What existed only in the artist's consciousness becomes publicly available, capable of being encountered by others. Spirit becomes flesh. The intangible becomes tangible. And this making, L'Engle insists, carries "the breath of life": the animating presence that distinguishes a living thing from a dead one, an artwork from a mere artifact.

L'Engle's understanding of the artist as "namer" deepens this vision. In the Genesis narrative, God creates the animals but brings them to Adam to name. The naming is not incidental to creation but participates in it. To name truly is to recognize the essential nature of a thing and to call that nature forth, giving it fuller existence through the act of articulation. The artist is not creating ex nihilo as only God can do. But the artist is "co-creating," working alongside the ongoing creative activity of God, bringing cosmos from chaos.¹²

This is why L'Engle insisted that "in art, we are, ourselves, being created."¹³ The making changes the maker. The encounter with the work changes the viewer. Art is not merely production and consumption but transformation, a process in which all participants are shaped by their participation.

Consider what this means concretely. When a painter spends months working on a single canvas, living with it daily, waking in the night with new ideas about how to solve a particular problem, the painting becomes a record of that lived time. The viewer may not consciously perceive all this, but something of it communicates nonetheless. We sense the care, the struggle, the moments of breakthrough and the longer stretches of uncertainty. The painting is not just a visual arrangement but a deposit of lived experience, and it is this deposit that carries what L'Engle called "the breath of life."¹⁴

If this seems mystical, consider its negative. We have all encountered works that are technically proficient but somehow empty. The Hollywood blockbuster made by committee, assembled from market-tested elements, engineered to maximize opening weekend returns: it may be entertaining, but it does not nourish. The corporate art in hotel lobbies, designed to offend no one and thus to mean nothing: we look through it rather than at it. These are not failures of technique. They are absences of inner need, vacuums where the breath of life should be.¹⁵

4. The Space Between: What AI Does and Does Not Do

We have now assembled the conceptual tools necessary to think carefully about artificial intelligence and artistic making. The question is not whether AI can produce outputs that resemble art, the question is whether those outputs participate in the incarnational pattern we have been describing, whether they carry the breath of life, whether they spring from inner need, whether they reach asymptotically toward meaning or merely simulate such reaching.

To answer this question, we must first understand what AI actually does, stripped of both the inflated claims of its enthusiasts and the dismissive caricatures of its critics.

Large language models like ChatGPT or Claude operate by predicting the most probable next token in a sequence. Trained on a vast corpus of human-generated text, they learn statistical patterns: which words tend to follow which other words, which sentence structures are common, which ideas cluster together in human discourse. When given a prompt, they generate responses by repeatedly selecting the token that is most probable given everything that has come before. The process is sophisticated, involving billions of parameters and attention mechanisms, corrections, and rewards, but it is fundamentally statistical. The model does not "understand" in any sense we would recognize; it calculates.

Image generators like DALL-E, Midjourney, or Stable Diffusion operate by similar logic: they learn statistical relationships between verbal descriptions and visual features from millions of image-text pairs, then synthesize new images that correspond to a given prompt. The results can be technically accomplished, visually striking, sometimes genuinely beautiful by conventional measures.

Notice what is absent from both processes. There is no inner need, no innere Notwendigkeit. The model does not generate an image because something within demands expression. There is no asymptotic reaching toward meaning, because the model has no meaning to reach toward. There is no incarnation of spirit in flesh, because there is no spirit to incarnate. The model produces tokens according to statistical patterns learned from human-generated training data.

This is not a criticism of AI. It is simply a description of what AI is. A hammer is not deficient because it cannot see; seeing is not what hammers do. An AI model is not deficient because it lacks inner need and lived experience; having inner need and lived experience is not what AI models do. The confusion arises only when we mistake the outputs for something they are not, when we imagine that plausible generation equals meaningful expression.

The difference matters because it determines what we are actually encountering when we engage with AI-generated content. When we stand before a Rembrandt, we encounter the residue of a consciousness grappling with mortality, dignity, suffering, and grace. Something in the work speaks to something in us, and this speaking presupposes a speaker with something to say. When we stand before an AI-generated image in the style of Rembrandt, we encounter a statistical artifact, an arrangement of pixels that corresponds to patterns the model learned from studying actual Rembrandts. The meaning we find in it is meaning we bring to it, projected onto a surface that holds nothing of its own.

This projection is not nothing. Human beings are meaning-making creatures; we find significance in cloud formations and inkblots and the cracks in old walls. But there is a difference between discovering meaning and receiving it, between projecting significance onto a blank surface and encountering significance that has been deposited there by another consciousness.

We must be precise here, because the point is subtle. All art requires what we might call participatory activation: the viewer's engagement, their willingness to meet the work, the meanings they bring that allow the work's meanings to kindle. A painting does not interpret itself; it requires a consciousness to complete the circuit. But participatory activation is not the whole story. In human-made art, there is also what we might call deposited presence: the residue of the maker's consciousness, their struggle, their inner need, embedded in the work itself. The viewer's participation activates something that is genuinely there. In AI-generated content, participatory activation has nothing to activate except the viewer's own projections. The surface receives meaning; it does not offer it. Both processes involve the viewer, but only one involves an encounter between consciousnesses.

5. The Junkspace of the Mind

The Dutch architect Rem Koolhaas coined the term "junkspace" to describe a particular quality of contemporary built environments: the airport terminal, the shopping mall, the convention center.¹⁶ Junkspace is not ugly, exactly. It is beyond aesthetic judgment, neither beautiful nor ugly but merely functional, optimized, relentlessly adequate. It is space from which particularity has been designed out, leaving only generic smoothness. Junkspace, Koolhaas argued, is what architecture becomes when it abandons any aspiration beyond accommodation. This is the point when it gives up on meaning and settles for flow.¹⁷

The concept extends unnervingly well to AI-generated content. If we imagine an aesthetic equivalent of junkspace (call it "junkart" for the sake of argument) we have something very close to what current AI systems produce. Not ugly, exactly. Not incompetent. It is nonetheless relentlessly adequate, optimized for smooth reception, engineered to minimize friction between production and consumption. AI-generated images tend toward a certain slick competence that looks impressive on first glance but dissolves under sustained attention. There is no there there, no depth to discover, no rough edge to catch on.

This is not accidental. It is a structural consequence of how AI generation works. The models learn what is most common, most expected, most statistically average in their training data. When they generate, they produce outputs that hew toward the center of the distribution, outputs that are, in a precise mathematical sense, generic. Unusual combinations, idiosyncratic choices, the particular strangeness that marks a genuine artistic vision: these are by definition unlikely, and systems that maximize likelihood systematically exclude them. This is not to say that AI systems never produce surprising outputs. They do, occasionally, generate images or texts that exceed mere adequacy.18 But these exceptions do not refute the systemic tendency; they are statistical outliers in a process that structurally favors the mean. The critique is ecological rather than absolute. The question is not whether AI can ever produce something interesting but what happens to a culture when the default mode of image-making tends inexorably toward the generic.

Human creativity, by contrast, is often most valuable precisely when it departs from the mean. The breakthrough insight, the novel combination, the vision that sees what no one has seen before: these are by definition unlikely, and their unlikelihood is part of what makes them significant. The artistic tradition advances not by producing more of what already exists but by producing what does not yet exist, by pushing into territory that the distribution of prior work did not cover.

If AI-generated content comes to dominate our visual and verbal environments, we risk a kind of cultural regression to the mean. This risk emerges not because any individual AI output is harmful, but because the aggregate effect of millions of outputs all hewing toward the statistical center will reinforce that center, crowd out the strange and difficult work that might pull us in new directions, and gradually erode our capacity for aesthetic discernment. When we become accustomed to the smooth adequacy of junkart, we may lose the patience for art that demands more of us. We may begin to prefer the frictionless to the meaningful, the immediately pleasing to the slowly rewarding.

There is a sense in which AI-generated content functions like a virus. It is not malicious, but parasitic on and potentially destructive of its host. The training data on which AI models depend was created by human beings working from inner need, reaching asymptotically toward meaning, incarnating spirit in matter. The models extract patterns from this human labor and reproduce them without the inner life that generated them. They then flood the cultural environment with these extracted patterns, making it harder for new inner life to develop and express itself. The host weakens; the parasite proliferates; and we may not notice what we have lost until it is gone.

6. The Irreducible Question: Are We the Same or Different?

Here we arrive at the tension that cannot be resolved by argument alone but only acknowledged and inhabited. The question underlying everything we have said is whether human consciousness and AI processing are fundamentally different in kind or merely different in degree.

If you hold, as many neuroscientists and philosophers of mind do, that human consciousness is itself a kind of computation, that the brain is a biological computer and thinking is information processing, then the distinction we have been drawing may seem arbitrary. On this view, there is no "inner need" in any metaphysically robust sense, only patterns of neural activation that we describe using mentalistic vocabulary.

This computational view has sophisticated defenders. Daniel Dennett argues in Consciousness Explained that what we call consciousness is not a unified inner theater but a "multiple drafts" process, various neural subsystems competing for influence, with no central observer watching the show.¹⁹ Douglas Hofstadter, in I Am a Strange Loop, suggests that the self is a pattern, a "strange loop" of self-reference that emerges from sufficiently complex information processing and could, in principle, arise in any medium capable of supporting such loops.²⁰ If they are right, then what Kandinsky called "inner need" might be nothing more than a particularly complex pattern of activation, and AI, given enough sophistication, might eventually produce something functionally equivalent.

The argument deserves serious engagement. But it faces a difficulty that its proponents have not fully resolved: the problem of phenomenal consciousness, what philosopher David Chalmers calls "the hard problem."²¹ Even if we grant that the brain processes information, that neural patterns compete and resolve, that selves are strange loops of self-reference, we have not thereby explained why there is something it is like to be a conscious being. We have not explained why the processing is accompanied by subjective experience at all. Thomas Nagel's famous question remains: we can describe all the functional properties of a bat's sonar system, but we cannot thereby know what it is like to be a bat.²² The explanatory gap persists. John Searle's Chinese Room thought experiment makes a related point: a system can manipulate symbols according to rules without understanding what those symbols mean.²³ Syntax is not semantics. Computation is not comprehension.

None of this proves that consciousness is irreducible: proof may be impossible in either direction. But it does suggest that the computational view has not yet earned the confidence its proponents sometimes display. The hard problem remains hard. And until it is solved, the possibility that human consciousness involves something beyond computation, something that matters for art, cannot be dismissed as mere mystification.

If, however, consciousness is not reducible to computation—if there is something it is like to be a human being that functional descriptions cannot capture—then the distinction holds. Art made by ensouled beings carries something that art made by machines cannot carry. Simulation is not the same as the real thing.

We cannot adjudicate this dispute here. It involves the deepest questions in philosophy of mind, questions that have occupied thinkers for centuries without resolution. What we can do is note that the position one takes on this underlying question will largely determine one's evaluation of AI-generated content. Those who see consciousness as computation will see AI art as continuous with human art. Those who see consciousness as irreducible will see AI art as fundamentally different, however convincing its surfaces.

This is the tension we must learn to inhabit. We do not have the luxury of certainty. We must make decisions about how to engage with AI-generated content in the absence of definitive answers to the questions that would ground those decisions. Do we use AI image generators in worship contexts? Do we allow AI to compose the music that accompanies our prayers? Do we read AI-generated devotional content as if it came from a human consciousness grappling with faith? These questions press upon us now, and we cannot defer them until the philosophers have settled the mind-body problem, if ever they do.

7. What Embodied Practices Reveal

The arguments we have made thus far about inner need, incarnational making, and the breath of life might seem abstract, even mystified, to readers trained in analytical modes of thought. On closer examination, they find concrete grounding in what we might call embodied practice: the making that happens when human bodies engage physical materials in real space and time. This dimension of artistic creation has received sustained attention from phenomenologists, particularly Maurice Merleau-Ponty, who argued that human knowledge is fundamentally incarnate—that we know the world not primarily through detached cognition but through our bodily engagement with it.24 The painter does not first conceive an image mentally and then execute it manually; the thinking happens in the hand, in the encounter between flesh and pigment and surface. The knowledge that emerges from such encounters is what philosopher Michael Polanyi called "tacit knowledge" or knowing that exceeds what can be explicitly articulated or formalized.25

This has profound implications for our question about AI and art. If artistic knowledge is fundamentally embodied, then systems without bodies cannot possess it. They can simulate its outputs but not replicate its processes. The distinction is not merely technical but ontological.

Consider what happens when a photographer waits for light. The body registers changes no sensor can measure: the quality of stillness before a cloud passes, the particular weight of late afternoon. These embodied perceptions inform decisions that shape the final image in ways the photographer herself may not be able to articulate. The resulting photograph carries traces of that waiting, that presence, that slow attention.

When a sculptor's hands meet material, a dialogue begins that has no computational equivalent. The material resists; it has properties and preferences; it does not accept every intention equally. The sculptor must negotiate, must discover what the material will and will not do, and this negotiation becomes part of the work. The finished sculpture is not just shaped matter but the residue of a relationship between maker and medium.

When a semiotician examines the gap between sign and signified, they encounter the asymptotic condition made visible in concrete instances. The word is not the thing; the image is not the object; the map is not the territory. AI systems operate entirely within the realm of signs, arranging signifiers according to learned patterns, with no access to or care for what those tokens refer to. The meaning of those signifiers, if meaning there be, comes from elsewhere.

When a phenomenologist attends to the sheer thereness of a material object, they engage a practice that depends on having a body, on being located in space. AI has no body. It encounters no objects. Whatever it does with representations, it does not encounter them. Encounter is not the same as processing.

These practices matter because they ground abstract questions in concrete realities. It is one thing to argue philosophically about inner need and incarnation and the breath of life. It is another to feel clay yield under your hands, to sense light shift across an afternoon, to know in your body the moment a composition resolves. Artists who make with their hands do not merely think about making; they make, and in making they know what theory cannot fully articulate. This knowledge, pre-reflective, tacit, irreducibly particular, speaks where theory falls silent.

8. Living in the Tension

We have traced a path from the asymptotic fidelity of language through incarnational theories of art to the challenge posed by AI generation. We have tried to be honest about the difficulty of the questions, about the ways our answers depend on prior commitments about consciousness and meaning that we cannot prove. We have suggested that AI-generated content, whatever its qualities in particular instances, tends systematically toward a kind of adequate emptiness, a junkspace of the mind that may gradually erode our capacity for aesthetic discernment and spiritual depth.

But we have not resolved the tension, and we do not claim to resolve it here. The questions remain: Is there something in human consciousness that exceeds computation? Is inner need a real property of mind or a folk-psychological fiction? Does art carry something from maker to viewer, or do we merely project significance onto surfaces that hold nothing of their own? Can AI-generated religious content be spiritually nourishing, or is it necessarily empty, a simulation of devotion that leaves the soul unfed?

These questions will not be answered definitively by panels or papers. They may not be answerable at all in the way we typically want answers: clear, decisive, and binding. What we can do is hold them open, resist the premature closure that would let us stop thinking, and attend carefully to what our practices reveal. The artists who make with their hands, who encounter resistance and limitation, who work from inner need toward uncertain expression: they know something. The traditions that have reflected for millennia on the relationship between spirit and matter, between the visible and the invisible: they know something too. We need both kinds of knowing, and we need the humility to admit what we do not and perhaps cannot know.

The stakes, however, are real. If we are wrong about consciousness, if it is reducible to computation and AI is therefore continuous with human creativity, then our concerns about AI art are misplaced, and we should embrace the new tools as extensions of human capacity. But if we are right, if consciousness involves something that computation cannot capture and art requires something that AI cannot provide, then we face a genuine spiritual danger. The danger is not that AI will create great evil but that it will create endless adequate mediocrity, drowning the signal of genuine expression in noise, eroding our capacity to recognize the difference between the living and the merely lifelike.

We do not know which of these futures we inhabit. We will only find out by living forward, by paying attention, by trusting our encounters with genuine art to teach us what no theory can fully convey. The panel that accompanies this paper is an exercise in such attention: four artists reflecting on what their embodied practices reveal about the nature of making, the necessity of presence, the irreducibility of the encounter between maker and material. We offer it not as a set of answers but as a demonstration of the questions, an invitation to dwell in the tension rather than escape it.

If we cannot offer certainties, we can at least suggest a posture. The posture is one of deliberate slowness in a culture of instantaneous generation, of patient attention in an environment of endless content, of embodied presence in a world increasingly mediated by screens. For artists, this means continuing to make from inner need even when the market floods with adequate alternatives, trusting that the breath of life, however difficult to name, remains recognizable to those who have not lost the capacity to perceive it. For communities of faith, this means cultivating that capacity: teaching ourselves and our children to sit with difficult art, to wait for meaning rather than demanding immediate payoff, to value the struggle encoded in handmade things. For educators, this means resisting the pressure to treat AI-generated content as equivalent to human expression, not from technological anxiety but from commitment to the formation of souls who can tell the difference.

None of this is a program. It is closer to a discipline—the discipline of those who believe that consciousness is not computation, that spirit can inhabit matter, and that the asymptotic reach toward meaning is itself a form of prayer. We may be wrong. But if we are right, the discipline will have preserved something essential.

Notes

1. I first encountered this phrase on The Liturgists Podcast, episode 2, "God," originally aired February 11, 2014, in a conversation between Michael Gungor and Mike McHargue discussing the origins and application of language. I have adapted it here.

2. Wassily Kandinsky, Concerning the Spiritual in Art, trans. M.T.H. Sadler (New York: Dover, 1977).

3. Ibid., 24.

4. Ibid., 54.

5. Kandinsky's color theory assigned specific spiritual properties to each hue: yellow is aggressive and earthly, pressing outward toward the viewer with an almost physical force; blue is contemplative and heavenly, drawing the eye inward and suggesting infinite depth; green is "the bourgeois color," balanced and complacent, lacking the spiritual tension that produces real movement; red pulses with life, glowing from within. While these specific associations have not aged universally, the underlying claim that color communicates prior to and apart from rational interpretation remains compelling. See Kandinsky, Concerning the Spiritual, 36–45.

6. Makoto Fujimura, Culture Care: Reconnecting with Beauty for Our Common Life (Downers Grove, IL: InterVarsity Press, 2017); Makoto Fujimura, Art and Faith: A Theology of Making (New Haven: Yale University Press, 2020).

7. Fujimura, Culture Care, 22–24.

8. Fujimura, Art and Faith, 47.

9. George Steiner, Real Presences (Chicago: University of Chicago Press, 1989), 4.

10. Madeleine L'Engle, Walking on Water: Reflections on Faith and Art (Wheaton, IL: Harold Shaw, 1980).

11. Ibid., 17.

12. Ibid., 26.

13. Ibid., 149.

14. L'Engle is also clear on the inverse: "Bad art is bad religion, no matter how pious the subject." L'Engle, Walking on Water, 5.

15. A clarification is necessary here. The examples just given demonstrate that human-made work can fail to carry the breath of life. Inner need is not guaranteed by human origin. But the asymmetry runs one direction only: human art can fail to be incarnational, but AI art cannot succeed at being incarnational. The human artist who produces cynical or formulaic work has betrayed a capacity they possess. The AI system that produces similar work has not betrayed anything; it has simply done what it does. The difference is the difference between a person who refuses to speak and a recording that has nothing to say. Both may produce silence, but only one of them could have spoken.

16. Rem Koolhaas, "Junkspace," October 100 (Spring 2002): 175–90.

17. Ibid., 175.

18. These outputs are increasingly convincing, particularly to viewers who lack training in visual discernment or who, for various reasons, prefer not to scrutinize what they consume. The implications for education are significant. Critical thinking, long understood as the capacity to evaluate arguments and evidence, must now expand to include visual and textual literacy capable of distinguishing human expression from sophisticated pattern-matching.

19. Daniel C. Dennett, Consciousness Explained (Boston: Little, Brown and Company, 1991), 111–138.

20. Douglas R. Hofstadter, I Am a Strange Loop (New York: Basic Books, 2007), 84-87.

21. David J. Chalmers, The Conscious Mind: In Search of a Fundamental Theory (New York: Oxford University Press, 1996), xii–xiii.

22. Thomas Nagel, "What Is It Like to Be a Bat?," The Philosophical Review 83, no. 4 (October 1974): 439.

23. John R. Searle, "Minds, Brains, and Programs," Behavioral and Brain Sciences 3, no. 3 (1980): 417–424.

24. Maurice Merleau-Ponty, Phenomenology of Perception, trans. Colin Smith (London: Routledge, 1962), 251–276.

25. Michael Polanyi, The Tacit Dimension (Chicago: University of Chicago Press, 1966), 4.

Bibliography

Chalmers, David J. The Conscious Mind: In Search of a Fundamental Theory. New York: Oxford University Press, 1996.

Dennett, Daniel C. Consciousness Explained. Boston: Little, Brown and Company, 1991.

Fujimura, Makoto. Art and Faith: A Theology of Making. New Haven: Yale University Press, 2020.

Fujimura, Makoto. Culture Care: Reconnecting with Beauty for Our Common Life. Downers Grove, IL: InterVarsity Press, 2017.

Hofstadter, Douglas R. I Am a Strange Loop. New York: Basic Books, 2007.

Kandinsky, Wassily. Concerning the Spiritual in Art. Translated by M.T.H. Sadler. New York: Dover, 1977.

Koolhaas, Rem. "Junkspace." October 100 (Spring 2002): 175–90.

L'Engle, Madeleine. Walking on Water: Reflections on Faith and Art. Wheaton, IL: Harold Shaw, 1980.

Nagel, Thomas. "What Is It Like to Be a Bat?" The Philosophical Review 83, no. 4 (October 1974): 435–450.

Searle, John R. "Minds, Brains, and Programs." Behavioral and Brain Sciences 3, no. 3 (1980): 417–457.

Steiner, George. Real Presences. Chicago: University of Chicago Press, 1989.