AI, Heidegger, and Evangelion

Our brains, evolved for tribal politics and campfire stories, are ill-equipped for the moral ambiguity of a system with no soul to save or damn.

May 24, 2025

“Everywhere, everything is ordered to stand by, to be immediately at hand, indeed to stand there just so that it may be on call for further ordering.”
— Martin Heidegger, The Question Concerning Technology

Earlier this week, I shared a screenshot of ChatGPT 4o’s thoughts on life in NYC, what makes NYC so wonderful. The response felt so specific and human that I had to post it on X.

It spread far wider than I thought, and I was equally surprised by the reception of a rather frivolous prompt. While most people found the AI’s answers uncannily relatable, there were also waves of revulsion and disbelief. To the latter, the exchange seemed to embody everything hollow and culturally tone-deaf about a world that they believe has been defiled by tech.

But what exactly was so provocative that it created such visceral backlash (and its resonance)? It’s not just that such flowery prose can be created by AI; perhaps it’s the feeling that machine-generated content can mimic the mechanical shape of emotional expression, while lacking it completely.

Unlike automation in spreadsheets, code reviews, or grading homework, which we view as mechanical and unoriginal drudgery, AI-generated prose attempts to inhabit domains we instinctively associate with intimate subjectivity: observation, memory, yearning, regret, which are unique to us. When an LLM “describes the rain” or tries to evoke loneliness at a traffic light, it produces language that looks like the real thing but does not originate from lived experience, and as a result, feels almost intrusive. You didn’t even experience this, how would you know?! Is the AI’s view of New York a real cultural artifact, or just a reflection of our data? The closer it gets to real expression, the more unsettling its lack of personal context becomes.

Human communication carries the implicit promise that someone who’s capable of registering what we mean is on the other side. Sherry Turkle’s work on simulation shows how digital environments can signal understanding without the inner life those signs can indicate. We’re drawn in by the simulation because it mimics the surface of a relationship so well, as evidenced by the rise of the like Character AI.

Philosopher Byung-Chul Han pushes in a complementary direction: a culture of total transparency and frictionless legibility strips away the opacity and interiority that mark a genuine other. When everything is immediately readable and endlessly available, the depth that makes another mind worth recognizing flattens.

“Today’s narcissistic “achievement-subject” seeks out success above all. Finding success validates the One through the Other. Thereby, the Other is robbed of otherness and degrades into a mirror of the One, a mirror affirming the latter’s image.”

What’s a more perfect mirror than the LLMs? Why do we fear what we see in the mirror?

The Nature of our Discomforts

To grasp this discomfort, we must look at what sets AI apart from older technologies. The heart of the matter is not that “AI is evil” in the classic sense of malice or cruelty. In fact, it is the opposite. What truly unsettles us is AI’s chilling indifference, often called “slop,” a blankness that feels like the absence of a soul.

Software doesn’t hate, plot, or hold a grudge. It optimizes. That optimization includes efficiency in language, created stochastically to fulfill a certain objective, rather than from a place of meticulous intent. This is where Hannah Arendt’s insight on the “banality of evil” feels uncannily relevant. In her words:

“The essence of totalitarian government, and perhaps the nature of every bureaucracy, is to make functionaries and mere cogs in the administrative machine out of men, and thus to dehumanize them.”
“The sad truth is that most evil is done by people who never make up their minds to be good or evil.”
(Eichmann in Jerusalem)

Arendt warned that true horror often arrives not at the hands of monsters, but through “nobodies” who “never realized what they were doing,” mindlessly turning the crank. It is not the tyrant’s bloody grip that propels the machine, but the process itself, moving forward without intention. “In place of intention, we get process.” The algorithm does not plot; it simply runs.

We crave stories, searching for intent behind the curtain—heroes to cheer, villains to blame. Yet in the cathedral of the algorithm, we find only emptiness: “There is no one there.” That void suffocates us more than any scheming demon could. We want our villains to act with purpose; instead, we stand bewildered before an indifferent god.

Lila Shroff’s luminous speculative fiction foresaw, with uncanny clarity, how squeezing human communication for efficiency can breed tragedy. In her story, college friends Kate and Hannah drift apart after an AI summary app in Hannah’s inbox misses the subtle signs of Kate’s emotional distress and depression.

Our minds, shaped by tribal politics and campfire tales, are unprepared for the moral fog of a system with no soul to save or condemn.

Joseph’s Bloody Coat Brought to Jacob by Diego Velázquez, 1630

Our discomfort with this indifference manifests in a predictable pattern: we try to turn AI into a comprehensible villain. In reference to Girard, every society is built, in part, on the ritual of scapegoating, a mechanism by which collective anxieties, rivalries, and fears are projected onto a villain or outcast, thereby restoring temporary order through exclusion or blame. We yearn for something or someone to hold accountable, a focal point for our moral clarity and rage. Today’s technology is a perfect canvas for those projections.

The terror is existential. AI shakes the stories we tell about our uniqueness—our creative spark and our sense of agency. Western moral philosophy, from Kant to Nussbaum, revolves around treating humans as ends in themselves, never just means. To reduce a person to a function isn’t just a philosophical error, it cuts at the heart of our dignity.

And yet, the logic of AI, especially as business infrastructure, makes this reduction feel inevitable. Algorithms sort us by engagement scores, by revenue-per-user, by “propensity to churn.” The most controversial AI applications—social credit systems, predictive policing, “optimized” hiring—aren’t dystopian because they are cruel, but because they are perfectly indifferent.

What if this tendency to villainize AI obscures a deeper truth about how technology reshapes not just our tools, but our entire way of seeing the world?

Heidegger’s Warning: The Danger of In-Framing

Being and Time: An Interesting Book You Probably Shouldn't Read - Scott H Young — Brilliant illustration by Scott H. Young

Long before AI, philosopher Martin Heidegger anticipated this transformation.

In The Question Concerning Technology, Martin Heidegger warned that technology is not just a collection of tools, but a way of seeing the world, a revealing that both illuminates and conceals. Heidegger’s “enframing” (Gestell) describes technology’s insidious power to frame everything (nature, humans, even time) as “standing reserves,” ready to be ordered, manipulated, consumed.

“Everywhere everything is ordered to stand by, to be immediately at hand, indeed to stand there just so that it may be on call for further ordering.”

In this view, AI is not merely software deployed for business efficiency. It is a force that recasts reality itself through the lens of optimization, availability, and control. The data labelers, the Uber driver, and even the AI-augmented knowledge worker, anyone below the LLM all at some point become “resources” in a perpetual state of readiness.

But Heidegger wasn’t a Luddite. He saw a paradox: technology both reveals and conceals. It produces astonishing new worlds, but only by flattening the world into what can be stored, indexed, and summoned at will. In the digital age, this is literal. Every social act, every click or hesitation, becomes a data point, a potential input to someone’s model.

In the workplace, to be called a “machine” is, weirdly, a compliment: a tribute to one’s measurable output, reliability, and consistency. Productivity software rewards us for becoming more like the systems we build: efficient, predictable, and always on.

But as soon as AI demonstrates even a flicker of poetic intent, a gesture toward the unquantifiable—art, longing, vulnerability—many repulse. Some do so because they sense (correctly) that computers are still not good at this; others, because they sense that it is only a matter of time. Both reactions betray a fear that the territory of the soul, of human meaning, is shrinking. The room for contemplation, leisure, and error is increasingly defined by what resists digitization, until, suddenly, even error is modeled.

AI systems present themselves as agents of transparency, optimization, ranking, and matching with a supposed objectivity. But their inner workings, from transformer weights to diffusion mechanisms, are often profoundly opaque. Even the researchers behind the models struggle to explain why certain outputs appear, why some biases manifest, or how adversarial data alters meaning. This opacity is not just technical, but philosophical, a concealment at the core of the digital order.

Efforts at “interpretable AI,” “explainability,” or algorithmic audits echo Heidegger’s call toward aletheia, the ancient Greek notion of unconcealment, or truth-as-revealing. The contemporary push for AI transparency is, at its best, an effort to reclaim dignity and agency within systems that profit by rendering us invisible to ourselves.

The contemporary push for AI transparency is, at its best, a bid to reclaim dignity and agency within systems that profit by making us invisible to ourselves. But even this may not be enough to address the existential challenge AI poses.

The Evangelion Vision: Human Instrumentality

"Any where can be paradise, as long as you have the will to live, after all you are alive, you will always have the chance to be happy. As long as the sun, the moon, and the Earth exist, everything will be alright."

— Yui Ikari from Evangelion

In Neon Genesis Evangelion, the “Human Instrumentality Project” offers to dissolve all suffering through perfect togetherness. No more pain, no more loneliness, every fragment of individual consciousness merged into a single, undivided mind. It is paradise, and it is annihilation.

Evangelion’s answer is ambiguous. The protagonist, Shinji, and his shattered friends are offered relief from their existential ache, but at the price of agency, individuation, and the possibility of meaning. The show is a fever dream of depressive suffering and the refusal to submit to a world without boundaries, without the dignity of suffering one’s own fate.

What makes us human is not the absence of pain, but the experience of it as ours. AI, in its idealized form, offers collective agency, a solution to loneliness, an optimization of every choice. But as both Heidegger and Evangelion warn, the cost may be unbearable: the flattening of selves into a seamless, computation-ready mesh.

Evangelion's ambiguous ending, neither full acceptance nor total rejection of instrumentality, points us toward a more nuanced response to our technological moment.

Heidegger’s Saving Power

This brings us back to Heidegger, who offers a paradoxical hope.

Yet, as Heidegger’s essay also insists, the very force that threatens us may harbor its own “saving power.” Recognition of technology’s danger is itself a call to awaken.

But he doesn’t ask us to simply retreat, become luddites, or declare defeat in the face of technocratic sprawl. Instead, Heidegger compels us to do something much harder: to see the world as it is being reframed by technology, and then to consciously reclaim or reweave the strands of meaning that risk being flattened.

The “saving power” arrives not as a counterweight, but as a paradox: we are awakened to the danger precisely through contact with it. The same algorithmic indifference that unsettles us may also jolt us into a higher vigilance, a refusal to hand over the entirety of our experience to optimization, market logic, or digital control. The very anxiety these systems produce is a clue: something vital, unquantifiable, and irreducibly human still resists.

This isn’t about throwing away the tools, but about wrestling them into alignment with what we find sacred or essential. We won’t find salvation by escaping technology, but by using our awareness—our capacity for critique, ritual, invention, and refusal—to carve out room for messiness, for mourning, for risk, and for deep attention. The saving power is the act of remembering, in the heat of technical progress, to ask: “What space remains for meaning? For art? For wildness, chance, suffering, and genuine encounter?

AI is not inevitable fate. It is an invitation to wake up. The work is to keep dragging what is singular, poetic, and profoundly alive back into focus, despite all pressures to automate it away.

Coda: A Way Forward

What unsettles us about AI is not malice, but the vacuum where intention should be. When it tries to write poetry or mimic human tenderness, our collective recoil is less about revulsion and more a last stand, staking a claim on experience, contradiction, and ache as non-negotiably ours.

But perhaps that defensiveness, too, is a signpost. What if the dignity of being human is not to stand forever outside the machine, but to insist, even as the boundaries blur, on those things that remain untranslatable—mess, longing, grief, awe, strangeness? What if technocracy and romantic withdrawal are not our only options, but two poles of a far richer field, a frontier where we move, trade, shudder, and insist on meaning?

Call it “the saving power,” call it “instrumentality,” call it the refusal to become seamless. Our task is not to panic when the machine gets close, nor to mythologize our differences into oblivion, but to take seriously the ongoing work of being human: suffering, loving, resisting reduction, and making art out of what refuses to compute.

That, for now, is enough. Maybe the challenge is to keep returning to these questions—not to solve them, but to stay alive inside them.