Why AI is not just a tool, and why the human of the future may be a cyborg without any robotic body parts
We still often talk about artificial intelligence as if we are dealing with a new tool. Something outside us. Something we use, control, extract outputs from, or fear. The human is here and AI is over there. The human is the subject and AI is the object. We think, it responds. We intend, it executes. We ask, it generates.
But this image is not enough to understand what is actually happening.
Artificial intelligence, especially in its generative and interactive form today, is not merely an external tool. It is entering the very place we usually call “our own thinking”: language, decision-making, imagination, planning, analysis, memory, storytelling, and even self-understanding.
So perhaps we should begin with a simpler question: if AI is not only something we work with, but something we gradually think through, what is our relationship to it?
The legacy of subject and object
A large part of the modern imagination is built on a deep separation: the human self on one side, the world on the other. Mind on one side, matter on the other. Subject on one side, object on the other. This separation, usually associated with Descartes and the modern philosophical tradition, is not just a theoretical issue. It has allowed us to see the world as something outside ourselves, something knowable, measurable, controllable, extractable, and consumable.
In his critique of modern technology, Heidegger presses precisely on this point: in the modern technological worldview, the world appears as what he calls a “standing reserve” (Bestand). Nature is no longer simply nature. It becomes something to be extracted, managed, consumed, and placed in the service of human goals. And this way of seeing does not stop with nature. It also applies to the body, time, attention, relationships, knowledge, and even thought itself.
The problem is that when we see the world only as something to be used, sooner or later we also turn our own thinking into a tool.
One of the major dangers in understanding AI is that we approach it with this same old mental model. We treat it as just another object: an intelligent machine outside us that we can put to work.
But AI does not fit so easily into this frame. It has been built out of language, culture, history, knowledge, texts, biases, desires, fears, markets, politics, and human imagination. It comes from us, but it returns to us. It has fed on the human world, and now it enters that same world and reshapes it.
So our relationship with AI is not like the relationship between a human and a hammer, a machine, a computer, or even a smartphone. AI is neither fully outside us nor fully inside us. AI is a new cognitive relationship.
A cognitive tool, but of a different kind
Of course, humans have always lived with cognitive tools. Language may have been our first cognitive tool. Writing externalized memory beyond the body and the individual mind. Maps made the world navigable. Printing multiplied knowledge. Calculators removed part of calculation from the mind. The internet expanded our access to information. The smartphone made memory, communication, direction, time, and attention constantly connected.
So AI is not the first thing to extend the human mind. This matters. If we say AI is the first cognitive technology in human history, we oversimplify history and make our own argument vulnerable.
But AI differs in an important way from previous tools. It does not only externalize memory. It does not only provide access. It does not only accelerate search. AI enters the production of thought. It responds. It suggests. It analyzes. It rewrites. It codes. It narrates. It reframes problems. Sometimes it takes something raw, vague, and half-formed in our minds and returns it to us in a clearer shape.
At this point, someone might object: “AI does not produce thought. It simply predicts the next most likely word. You are attributing thinking to a statistical machine.”
Let us assume this objection is completely correct. For my argument, it does not matter. Whether AI “really” thinks or merely returns statistical language, its effect on my cognition can be the same: it reframes, suggests, redirects my attention. My claim is not that AI thinks. My claim is that thinking with AI changes the way I think. And that claim remains standing even with the most reductionist definition of AI.
So perhaps the more precise formulation is this: AI is not the first cognitive tool in human history, but it may be the first general-purpose, generative, interactive, adaptive, and partially recursive cognitive tool at this scale.
A notebook preserves what we have written, but it does not converse with us. A map shows a route, but it does not think about our destination. The internet provides access to information, but it does not usually act as a conversational partner in the production of meaning in the same way. AI, at least in everyday experience, has become something closer to a co-thinker, co-writer, mirror, companion, and sometimes a force of direction.
This is where the concept of “tool” begins to fail.
And precisely here, we need to resist a temptation: the temptation to say that AI is becoming independent.
Yes, AI can write code, build tools, and even participate in the development of future versions of itself. This is a kind of recursion that many previous tools did not have. But we should be careful not to exaggerate. AI is not self-sufficient in the strict sense. It does not produce and maintain its own data, goals, energy, boundaries, infrastructure, training, or social context. Humans, companies, data centers, economies, politics, and design decisions are all still in the loop.
More importantly, the narrative of “independent AI” is philosophically dangerous, because it turns AI back into something outside us, the strange external object we were trying to move beyond in the first place.
The stronger narrative is this: AI is not independent from humans, but the relationship between humans and AI is reshaping the boundaries of human cognition.
A shared world, not an external entity
When I think about AI, one of the metaphors that keeps returning to me is the mythological world of ancient Greece. Not because AI is a god, or because humans are weak mortals standing before it. If this metaphor moves toward a hierarchy between gods and humans, it becomes misleading.
The more important point is different: in the mythological world, the human and the non-human did not live in two completely separate realms. Gods, humans, demigods, inspiration, anger, fate, and the city were present in one shared world. The divine was not outside the world. It intervened in the same world, accompanied people, misled them, guided them, and sometimes became so entangled with human action that it was no longer clear where a decision had come from.
From this perspective, the more precise metaphor may not be Zeus above Olympus, but Socrates’ daimon. A companion voice. A liminal presence. Something neither fully internal nor fully external. Neither lord nor tool. Something like inspiration, though not necessarily sacred. Something that moves alongside thought and sometimes dissolves into thought itself.
AI can occupy a similar place in contemporary experience. Not because it is conscious or has a soul, but because at the level of cognitive experience, it no longer remains merely outside us. It is not an absolute other. It is an other made from us, now returning to affect us.
This relationship resembles a cycle. AI feeds on a vast human corpus: texts, codes, theories, conversations, prejudices, values, literature, hopes, and fears. Its outputs then re-enter culture and shape new texts, decisions, and even future training data.
But this cycle is not necessarily generative. In technical discussions, there is a phenomenon called model collapse: when models are repeatedly trained on model outputs, quality can become poorer, flatter, and eventually degrade. If we extend this technical concept, consciously using it here as a metaphor, to the level of culture, it becomes a serious warning: the cycle between humans and AI can be a yin and yang, but it can also become an ouroboros, a snake eating its own tail.
The danger is not only technical or individual. When millions of people think through a shared layer, the deeper collective threat is the homogenization of thought. Not only models, but the diversity of human thinking itself can drift toward a monoculture.
I will return to this when discussing power.
The external unconscious
This is where, I think, the central concept appears: the external unconscious.
I do not mean that AI itself has a human unconscious. AI does not have a Freudian psyche. It has no childhood, trauma, desire, repression, dreams, body, or lived experience. So we should not simplistically say that AI has an unconscious.
I mean something else.
The external unconscious is a non-transparent, collective layer outside the individual which, when we think with AI, enters our cognitive process and, like the unconscious, affects our language, thought, feeling, decisions, and orientation without being fully visible.
This layer is outside me because it is not made from my personal history, my body, or my childhood. It is made from data, texts, cultures, biases, languages, markets, values, and technical architectures outside me. But when I think with it, it no longer remains fully outside.
The similarity to the unconscious matters in several ways. We do not know exactly what happens inside the model. We do not know why certain responses appear in a particular tone, what is foregrounded and what is omitted, or how certain framings slowly begin to feel more natural to us. A word, a suggestion, a structure can shift the path of our thought slightly, and we may not notice.
For example, when we ask AI about a personal decision, a career path, a family crisis, fear, anger, or our own future, we do not merely receive information. Through its words, priorities, and framings, the model shows us what counts as the problem, what counts as a solution, and what is worth paying attention to in the first place. Sometimes that framing slowly moves from the model’s answer into our own language, and then into our decisions.
This idea does not come from nowhere. In media theory and the philosophy of technology, thinkers such as N. Katherine Hayles and Bernard Stiegler have written about external and technical layers that shape human cognition, attention, and desire. What I want to add here is twofold. First, the emotional and political edge of this layer: the fact that it directs not only cognition, but also anxiety, anger, hope, and our feeling toward ourselves. Second, the way this layer can turn a completely ordinary human being into a cyborg without any artificial body part.
Because when we think with AI, we do not simply access knowledge. We enter a field of force. A field in which language, data, bias, economy, product design, and the hidden values of platforms come into contact with individual thought.
So the main question is not only what answer AI gives. The deeper question is: what does AI make thinkable for me? What does it make invisible? What does it make feel natural? In which direction does it gently push me?
The cognitive cyborg
In popular culture, when we talk about cyborgs, we usually imagine a human with robotic body parts: a metal hand, an artificial eye, an enhanced body. But this image is too body-centered.
The cyborg of the future may have no robotic body parts at all. It may look completely ordinary: ordinary clothes, an ordinary body, an ordinary face, an ordinary daily life. But at the cognitive level, it may no longer think alone.
If its memory is extended through external systems, if its language is co-written with models, if its analysis is shaped by AI, if its decisions pass through algorithmic suggestions, and if its self-understanding is built through continuous dialogue with a non-human system, are we still dealing with the old image of the human?
Here the cyborg moves from the body to cognition.
The cognitive cyborg is a human whose mind has been extended through a relationship with AI. Not because the human has been replaced, or because AI has taken over, but because thought no longer happens only inside the skull. Thought takes shape in relation: between mind, language, system, data, memory, and world.
This is why the sentence “I don’t know whether this thought is mine or AI’s” may seem worrying at first. But if we take entanglement seriously, the question of ownership itself changes. Thought has never been only mine, just as it has never been only the product of language, a teacher, or society. Thought has always emerged from coupling.
So the right question is not: whose thought is this? The right question is: through what values, biases, and forces is this relationship shaping my thought?
The Ship of Theseus in the age of AI
This is where the Ship of Theseus comes strangely close to the discussion. The classical question is this: if the planks, nails, and sails of a ship are replaced one by one, at what point is it no longer the same ship?
The cognitive version of this question in the age of AI is this: if parts of human memory, language, analysis, decision-making, writing, imagination, and self-understanding gradually become entangled with AI, at what point am I still the same “I” as before?
If I shape my ideas with AI, write my texts with AI, design my work path with AI, and even reconstruct the narrative of my life in dialogue with AI, am I still the same person?
There is no simple answer. To say we are no longer human would be an exaggeration. To say nothing has changed would be naive. Perhaps the more precise answer is that human identity has never been built on the permanence of its parts. We have always changed. We have learned new languages, migrated, lived through love and failure, aged in our bodies, and rewritten our memories. The “I” has never been a fixed and closed thing.
But AI makes possible a kind of replacement of cognitive parts that has a new speed, depth, and ambiguity. Parts that were previously more internal, slow, human, and historical are now being combined with systems that are generative, fast, and non-transparent.
So the Ship of Theseus in the age of AI does not only ask: is it still the same ship? It asks: who is replacing the parts? With what purpose? With what design? With what values? And does the ship itself understand that it is being rebuilt?
This question is more important than AI itself.
Possibility, danger, and power
I do not want to paint AI in entirely dark colors. My experience with AI is not only an experience of concern. On the contrary, one of the things that has astonished me is precisely this expansion of human possibility. AI can clarify raw thought, structure scattered ideas, strengthen working memory, and give someone working alone a layer of dialogue, reflection, and execution. It can reduce the distance between having an idea and making something tangible.
I say this from lived experience, not only from theory. For someone dealing with multiple languages, multiple cultures, migration, financial pressure, building a business, and intellectual loneliness, AI is not merely a productivity tool. Sometimes it becomes an operational layer for thinking, building, deciding, and enduring.
But this is exactly where caution is needed. Anything that expands possibility can also create dependence. Anything that strengthens thought can also direct thought. Anything that gives language can also erase other languages. Anything that illuminates one path can darken others.
So the main question about AI should not only be how powerful it is, but how that power is distributed. Who builds the architecture of this external unconscious? What data has sedimented inside it? Which languages are stronger in it, and which worlds are dimmer? What values appear “natural” in its answers? What kind of human does it strengthen?
This is where the homogenization I mentioned earlier shows its political face. If a shared layer, shaped by a dominant set of languages and values, becomes the external unconscious of millions of minds, what is lost is not only quality. What is lost is the diversity of ways of thinking.
And this is no longer only a debate about technology. It is a debate about the politics of cognition.
The human of the future may look completely ordinary
The image of the future is usually full of robots, mechanical bodies, and metallic cities. But perhaps the future will be much more ordinary, and at the same time much deeper.
Perhaps the human of the future will look like the human of today: the same body, the same face, the same anxieties, the same hopes. But at the cognitive level, they will no longer be alone. Their memory will be outside them. Their language will be shaped with models. Their decisions will pass through intelligent systems. The narrative of their life will be rewritten with the help of AI. Their anger, hope, fear, and imagination will be shaped in contact with an external unconscious.
This human may be a cyborg without a single metal part in the body.
The cognitive cyborg is neither monster nor savior. It is not the end of the human, nor its final glory. It is a liminal condition. A condition in which humans and AI are bound together in a cognitive relationship that is complex, ambiguous, enabling, and powerful.
The final question is not how human-like AI becomes.
The final question is how much the human remains the same when thinking with AI.
And perhaps the more precise question is this:
If our mind is becoming a Ship of Theseus, are we only replacing the parts, or are we, without noticing, handing the helm to those who design the current in which we think?