AI Through Heidegger’s Question Concerning Technology
In recent years, artists have begun living through a new experience in the history of art. An illustrator discovers that anyone can type “in the style of” followed by their name and receive, within seconds, an image that looks as if it came from their hand. A novelist finds that a model knows the rhythm of her sentences. It is not one specific work that was taken; at least that could be named in a lawsuit. Something more elusive has been extracted: the way of seeing itself, the result of twenty years of work, turned into a resource that can be summoned by command.
The immediate response has been legal: lawsuits by artists and writers against AI companies are moving through the courts. But copyright is the shell around a deeper question. The phenomenon does not stop with artists. Platforms update terms of service to train models on our posts. Conversations with chatbots become training data. The face becomes a vector, attention becomes an engagement metric, personal style becomes weights in a model. Something in the world has changed: every human expression has become, by default, raw material.
The philosopher who gave the sharpest language to this kind of transformation wrote about dams and factories decades before the internet existed.
Technology Is Not a Tool
Martin Heidegger, the German philosopher, lived in the twentieth century, through rapid industrialization, two world wars, and the rise of technology from useful instrument to force organizing whole lives. He is also, it must be said plainly, one of the most difficult and compromised figures in modern philosophy: his support for Nazism stains his biography and requires that he be read with open eyes. And yet his 1954 essay “The Question Concerning Technology” remains one of the most important texts for our age precisely because it refuses the question everyone asks.
The usual question is whether technology is good or dangerous. That question assumes technology is a tool, a large hammer whose value depends on the hand that holds it. Heidegger argued that this assumption is itself the mistake. Technology is not merely a collection of tools added to an existing world. It is a way in which the world appears.
In the modern technological regime, he argued, everything is revealed in advance as available stock: a reserve of resources waiting for use. The river appears as hydroelectric potential, the forest as timber, the mountain as mineral deposit. Not because some evil person decided this, but because this has become the way of seeing itself, before any particular decision.
The danger, he said, is not that the machines malfunction or rebel. It is subtler: that this way of seeing becomes exclusive, and stops feeling like a way of seeing. That we forget the river once appeared as something else - not merely “also beautiful,” but disclosed differently at the root. And eventually we apply the same gaze to ourselves, appearing to one another as human resources, human capital, inventory.
The New Mine
Heidegger died in 1976, and the gaze he described continued exactly along the path he saw, inward. Every previous industrialization mined the outside: coal, oil, timber, water. The current movement mines something that was never before a resource: the residue of inner life. What people said to one another, drew, asked in the middle of the night, wrote and deleted. A language model is precisely this: the industrial distillation of billions of human acts of expression into one usable fluid. It is not done out of malice, and it could not be done otherwise. For a machine to learn something, that something must first be turned into data. Extraction is not a flaw of this technology. It is its condition of existence.
That is why the title question is phrased as it is. Not “what will AI do to us,” but what has it already made into a resource before we ask it anything. The answer is wider than expected: art, conversation, attention, and ultimately the mere fact that human beings express themselves somewhere.
But there is one strangeness in this movement that Heidegger did not know, and it may be the truly new thing. A coal mine does not produce coal. This system does. It feeds on culture in order to produce culture. Its outputs - text, images, answers - flow back into the same network from which it draws. Researchers already worry about model collapse: models unknowingly trained on the outputs of previous models and gradually degraded. This is the first technology that mines meaning, and the first that refills its own mine with waste.
If this logic continues, the scarce resource of the coming decades may not be compute. It may be the unextracted: human expression created in live encounter, not yet passed through the conveyor belt. Early signs are already visible. AI companies are signing licensing agreements with news agencies and publishers to secure high-quality human text. The market has begun pricing the shortage.
The Serious Objection
Now the other side, because Heidegger is too easy to become addicted to. He can invite romanticism: the peasant in the field, the Greek jug, the glowing pre-industrial world. It is easy to forget that world also included plagues, hunger, and infant mortality, and that the extracting gaze - the gaze that turns everything into data - is also the gaze that decodes genomes and develops medicines. Technology does not only conceal the world. It also reveals it, literally. And every generation has mourned authenticity. People said photography would kill painting; it gave birth to new art. Perhaps generative models are simply the photography of this generation, and artists are already doing things with them that were not possible before.
These objections are true, and Heidegger himself, surprisingly, would agree with part of them. He never argued that we should throw away the machines. The most famous line in his essay quotes a poem: where the danger is, there also grows the saving power. For Heidegger, salvation is not escape from technology but clear sight of it: recognizing that it is one powerful and legitimate way of revealing, and refusing to let it become the only one.
The problem is not that the machine sees everything as data. That is its work. The problem begins when we can no longer see otherwise. The danger is not that the machine will hate us. It is that we will get used to seeing like it.
What Remains Outside
The practical conclusion of this essay is therefore not boycott and not panic. It is the practice of distinction. A conversation with a friend is not potential data unless you choose to record it. A lesson a teacher gives, a drawing in a notebook, something written for one person - these are spaces in which the extracting gaze has not yet become the default, and the decision whether it will is still, at least partly, in our hands.
Heidegger asked of every technology: what does it let us see, and what does it hide while we are looking? In the version belonging to our age, the question becomes intensely personal and worth carrying in your pocket: what in your life do you refuse to let become data? And just as important: do you still know how to explain why?
Because on the day the answer becomes “actually, nothing,” the machine will not have won. There will simply be no difference left to lose.
Sources for further reading
- Martin Heidegger, “The Question Concerning Technology”
- Hubert Dreyfus, What Computers Still Can’t Do
- Shoshana Zuboff, The Age of Surveillance Capitalism