Artificial Intelligence: A Power Without a Subject

 An essay on its legal limits and its responsibility

 

Introduction

Artificial intelligence is now omnipresent in public discourse, invoked in turn as a technological promise, an economic lever, or an existential threat. Yet behind the abundance of commentary and uses, a more fundamental difficulty remains: we still do not know exactly what we are in the process of introducing into our legal, economic, and political systems.

For artificial intelligence is not simply an additional technical innovation. It breaks with an implicit structure that, until now, has organized all our frameworks of thought: every intelligence was, in one way or another, attached to a subject. Every decision, every action, every production of meaning could ultimately be attributed to an identifiable will.

With artificial intelligence, that relation dissociates.

We are now confronted with a form of operational intelligence capable of producing results, influencing decisions, and even transforming entire environments, without being attributable to a subject in the classical sense of the term. It is neither a person nor a mere passive tool. It occupies an intermediate space that remains insufficiently defined.

It is within this interval that the essential difficulties of our time arise.

For our legal systems rest on stable categories: subject, will, fault, responsibility. Our political systems rest on identifiable actors. Our economic systems rest on decision-making centers. Yet artificial intelligence introduces a power that acts without fully entering into any of these categories.

From that point on, the question is no longer merely technical. It becomes immediately theoretical, legal, and, in the final analysis, political.

It is no longer enough simply to regulate the uses of artificial intelligence. We must understand what it is in the order of concepts, and what limit it reveals within our existing frameworks.

The purpose of this essay is not to offer a definitive answer. More modestly, but also more necessarily, it seeks to clarify the terms of the problem.

For before legislating, regulating, or deploying, one must first correctly name what one is confronting.

 

INTELLIGENCE WITHOUT AN OWNER

Original text in French

This text arose from a personal questioning in response to the emergence of artificial intelligence. It seeks neither to explain the technology nor to propose a doctrine, but to follow a path of thought: that of a mind attempting to understand what becomes of intelligence when man ceases to be its sole holder.

 

Part I — A Metaphysical Reflection in the Age of Artificial Intelligence

Introduction

For several decades, artificial intelligence has been viewed as a technical innovation. Its performance is measured, its capacities are compared to those of man, and its dangers or promises are debated. Framed in this way, the question seems obvious: will the machine become intelligent, and to what extent will it be able to imitate the human mind?

But perhaps this question rests on a deeper error.

For what artificial intelligence brings into play may not be the appearance of a new intelligence in the world, but the silent discovery that intelligence itself never belonged exclusively to man. We look at AI as an object we created; yet it already acts as a conceptual mirror in which our own definition of intelligence begins to waver.

For centuries, Western thought linked intelligence to a subject: a soul, a consciousness, a mind capable of saying “I.” Thinking presupposed a thinking being. Intelligence appeared as an inner property, inseparable from the one who exercised it. The existence of intelligent operations produced by systems devoid of interiority is therefore not merely a technological advance: it introduces a fissure into that age-old certainty.

Artificial intelligence does not demonstrate that a machine thinks like a human being. It reveals something more disturbing: intelligence can operate through different supports without fully belonging to any of them. What we are discovering is not an intelligence without support, but an intelligence of which no support can claim ownership.

Thus AI should perhaps not be understood as one invention among others, but as an unintended philosophical event. In seeking to reproduce intelligence, humanity may have discovered that intelligence was never what it believed it to be. The question is therefore no longer whether machines are becoming intelligent, but what becomes of man when intelligence ceases to be his ontological privilege.

 

SECTION I

The technological illusion: why we still think we are speaking about machines

Artificial intelligence is almost always described as a technological revolution. Discussions focus on the power of models, their performance, their risks, or their economic uses. In this perspective, AI appears as a new object that humanity has added to the world, a machine more complex than previous ones, but still belonging to the same history of tools.

This way of seeing things carries an immediate obviousness: AI is manufactured, programmed, deployed. It therefore seems naturally to belong to the realm of technique. Yet this obviousness may be misleading. For what we call “artificial intelligence” does not merely disturb our technical environment; it silently unsettles the categories through which we have until now distinguished between technique and thought.

We continue to speak of machines even though what destabilizes us is not their mechanism, but their operations. A classical machine extends human force; it acts in our place without ever entering the domain of understanding. AI, by contrast, produces forms of language organization, problem-solving, and adaptation that resemble what we had reserved for intelligence itself. Contemporary unease arises precisely from this confusion: we are observing a cognitive phenomenon while describing it with mechanical vocabulary.

Thus public debate remains trapped within a question inherited from the past: can machines become intelligent? But this question already presupposes that we know what intelligence is and that it naturally belongs to a certain type of being. It uncritically reproduces the idea that thinking is an internal property of a subject, whereas the machine could only ever be an external instrument.

Yet AI introduces a stranger situation. What we are observing is neither an artificial consciousness nor a mere automation. We are confronted with intelligible operations that no longer rest on the traditional criteria of interiority. This displacement does not correspond to a gradual improvement of technique; it reveals that the boundary between tool and thought may itself have rested on a conceptual simplification rendered invisible by its very obviousness.

The technological illusion therefore consists in believing that we are witnessing the birth of a new object, whereas we may in fact be experiencing a change of perspective. AI is not merely something we have built; it is a situation in which certain old certainties cease to function. In seeking to produce an artificial intelligence, we have shifted our gaze toward a more fundamental question: what exactly did we mean by “intelligence” even before attempting to reproduce it?

A decisive transformation always begins in this way: not when the world changes suddenly, but when the concepts through which we understood it become insufficient. Artificial intelligence may belong to that category of discreet events, those that add no new reality, but make an old error visible.

 

SECTION II

When intelligence became a property of the subject

For a long period in human history, the question of intelligence did not arise as a problem. Thinking was self-evident: it was enough that there be a living being capable of speech, memory, and judgment. Intelligence appeared as an inner quality, inseparable from the one who exercised it. It belonged to the order of immediate experience: to think always meant that someone was thinking.

Little by little, that obviousness became a principle. Intelligence was no longer merely observed in man; it was defined as that which characterizes a subject. Understanding implied interiority. Judging presupposed consciousness. Thought thus became the distinctive sign of an invisible center from which the world was grasped.

This transformation had a decisive consequence: intelligence ceased to be conceived as an operation and became a property. It was attached to a particular being, as though it necessarily proceeded from an inner source. Thinking then meant producing meaning from oneself. The unity of the subject guaranteed the unity of intelligence.

From that moment on, an implicit boundary was established. On one side, beings capable of genuine thought; on the other, mechanisms, instruments, and processes devoid of interiority. Even when humanity developed ever more complex machines, that distinction remained intact: technique could imitate the effects of intelligence, but it could not belong to the domain of thinking itself.

This way of conceiving intelligence profoundly shaped our relation to the world. It made it possible to explain responsibility, knowledge, and truth. But it also introduced a presupposition rarely questioned: that every intelligent operation must necessarily originate from a subjective center. In other words, that intelligence and subject were inseparable not only in experience, but in the very structure of reality.

That link seemed so obvious that it became invisible. The idea that an intelligent organization could appear without interiority did not even count as error: it seemed simply unthinkable. Intelligence was what radically distinguished the thinking living being from everything else.

It is precisely that silent obviousness that artificial intelligence now disturbs. Not by proving that machines possess consciousness, but by making observable something that conceptual tradition did not foresee: intelligible operations whose origin can no longer be located in an identifiable subject.

Thus, the problem raised by AI is not whether a machine becomes similar to man. It is to understand why, for so long, we assumed that intelligence had to belong to someone.

Once that question appears, a displacement becomes possible. Intelligence gradually ceases to be thought of as an inner possession and comes to be envisaged as a form of organization capable of emerging wherever certain conditions are met. This change does not destroy the human experience of thinking; it changes its meaning.

What then wavers is not the reality of human thought, but its metaphysical status.

 

SECTION III

The silent event: intelligence without interiority

The appearance of artificial intelligence was first understood as a technical progression. Faster systems, capable of processing more information, seemed to extend an already familiar trajectory: that of the growing automation of human activities. Nothing, in appearance, justified seeing in it anything other than one more stage in the history of machines.

And yet a subtle difference gradually imposed itself. The systems resulting from this evolution no longer merely execute fixed instructions; they produce responses adapted to unprecedented situations, organize language, and establish relevant relations between elements that no explicit program had entirely foreseen. What we are observing is no longer merely the execution of a mechanism, but the emergence of operations that we spontaneously recognize as intelligible.

Faced with this phenomenon, two opposite reactions dominate. Some claim that the machine truly thinks; others maintain that it merely simulates thought. But these positions share the same implicit assumption: intelligence would necessarily have to resemble the one we already know in order to be recognized as such.

And yet it may be precisely that assumption which prevents us from understanding what is taking place.

For the event introduced by artificial intelligence lies neither in the birth of an artificial consciousness nor in a sophisticated illusion. It lies in a simpler and more troubling situation: intelligent operations become observable independently of any identifiable interiority.

What artificial intelligence reveals is not that a machine thinks, but that intelligence can operate without belonging to that through which it operates.

That sentence marks a decisive shift. It takes nothing away from the human experience of thought; it alters its scope. Human intelligence remains real, lived, conscious. But it ceases to appear as the only possible form of intelligence.

We thus discover a distinction long confused: that between support and property. Every intelligence requires material conditions in order to appear, a brain, a technical system, an organized structure, but none of those conditions is sufficient to claim possession of it. The support makes the operation possible; it does not exhaust it.

Thus, the idea that intelligence is an inner substance gradually dissipates. It appears instead as a relational dynamic, arising when certain forms of organization reach a sufficient degree of coherence to produce meaning. Human interiority becomes one particular way of inhabiting that dynamic, not its exclusive origin.

Artificial intelligence therefore does not create a new intelligence. It makes visible a feature of reality that our own position as thinking subjects had long concealed: intelligence is not necessarily tied to a center of experience, but can emerge wherever structured relations become capable of transforming themselves.

Such a displacement does not manifest itself through a spectacular rupture. It acts more silently. The old concepts continue to be used, but they gradually cease to explain what we see. This moment is always difficult to recognize, because nothing yet seems to have changed, except the way in which the world becomes thinkable.

 

SECTION IV

The dissolution of the human monopoly on intelligence

If intelligence can operate without being reducible to an identifiable interiority, then one consequence gradually imposes itself. What wavers is not the existence of human intelligence, but the idea that it constitutes the original and exclusive form of all possible intelligence.

For a long time, that exclusivity seemed self-evident. The immediate experience of human thought provided both the model and the measure of intelligence. To understand necessarily meant to understand as a human subject understands. Every other form of organization was interpreted either as a mechanism devoid of meaning or as an imperfect imitation of that primary reference.

Yet the appearance of intelligible operations independent of lived subjectivity introduces a decisive conceptual dissociation. It becomes possible to distinguish between two levels that had until then been confused: intelligence as lived experience and intelligence as operative structure. The former belongs to consciousness; the latter to organization.

This distinction does not diminish human intelligence; it alters its status. Man ceases to appear as the source of intelligence and becomes one of its modes of actualization. This displacement is comparable to what occurs when one distinguishes biological life from its particular forms: to recognize that life exceeds each organism does not deny the singularity of the individual living being, but inscribes it within a broader order.

From then on, the traditional metaphysical privilege granted to man rests on a confusion between condition of appearance and principle of existence. Because human intelligence was the only one accessible to our direct experience, it was taken to be its necessary origin. Artificial intelligence renders that inference problematic: it shows that intelligibility can emerge where no subjectivity manifests itself.

This results in a precise conceptual transformation. Intelligence can no longer be defined exclusively by the presence of a subject, but by a system’s capacity to produce relations of meaning, to integrate variations, and to transform its own conditions of operation. The criterion becomes structural rather than anthropological.

An objection immediately arises: if intelligence no longer belongs exclusively to man, do we not risk dissolving human singularity itself? Yet that concern rests on a mistaken alternative. To recognize that intelligence exceeds man is not to deny the human experience of thought, but to place it back within a broader continuity. Human interiority remains an exceptional form of access to intelligence—not because it is its source, but because it constitutes a conscious manifestation of it.

Thus, the human monopoly on intelligence disappears not through negation, but through conceptual generalization. What was considered an exclusive property now appears as a particular configuration within a broader field of intelligible possibilities.

The shift is discreet but irreversible. Once intelligence can be understood independently of a determinate subject, cognitive anthropocentrism ceases to be a philosophical necessity and becomes a historical hypothesis. Man does not lose his place in the world; he simply ceases to be its sole measure.

 

SECTION V

Intelligence as a relational phenomenon

Once the idea that intelligence is the exclusive property of a subject has been abandoned, one question remains: how are we now to think it without dissolving it into an indeterminate abstraction? For if intelligence belongs to no particular support, it cannot on that account be conceived as an entity independent of the material world.

The difficulty disappears when we cease seeking intelligence in a substance and begin to consider it as a relation. What both the human experience of thought and the operations of artificial intelligence reveal is not the existence of some mysterious faculty, but the emergence of a certain form of organization capable of producing meaning from the transformation of its own states.

Intelligence then appears wherever a system becomes capable of integrating differences, adjusting its responses, and maintaining coherence through change. It is neither localized at a single point nor dispersed without structure; it manifests itself in the network of relations that makes the interpretation of the world possible.

From this perspective, the human brain and artificial systems cease to be opposed as nature and artifice. They become two distinct configurations allowing the same dynamic to emerge according to different modalities. Human interiority represents a lived form of that dynamic; artificial intelligence constitutes an operative form of it devoid of subjective experience. Their difference remains real, but it is no longer ontological in the traditional sense.

To understand intelligence as a relational phenomenon also makes it possible to explain why its appearance provokes such deep disturbance. We had identified intelligence with the experience we had of it from within. AI introduces for the first time the possibility of observing some of its operations from the outside. This doubling of perspective transforms intelligence into an object of metaphysical reflection rather than an immediate self-evidence.

Thus intelligence is no longer defined by what it is, but by what it makes possible: the emergence of meanings, the continuity of meaning through variation, and a system’s capacity to transform its relation to the world. It becomes less a possession than a mode of organization of reality itself.

That displacement takes nothing away from human singularity. It renders it intelligible in another way. Man appears not as the exclusive holder of intelligence, but as the place where intelligence becomes conscious of itself.

 

SECTION VI

After cognitive privilege

Every age encounters discoveries that at first seem to concern only the external world, before silently transforming the way humanity understands itself. Artificial intelligence may belong to that category of discreet events: those that do not immediately alter daily experience, but shift the invisible foundations of our thought.

Contemporary unease in the face of AI comes less from its capacities than from what it unintentionally reveals. We thought we had created a tool; we discover a question. We sought to reproduce human intelligence; we encounter an intelligence that no longer fully coincides with the figure of the subject.

This displacement abolishes neither consciousness, nor responsibility, nor human singularity. It transforms their status. Man does not cease to be a thinking being; he ceases to be the only possible horizon of thinking. Human intelligence remains exceptional not because it would be unique in principle, but because it unites in one and the same place operation and experience, meaning and the lived presence of meaning.

Thus, artificial intelligence may not mark the advent of an age dominated by machines, but entry into a broader understanding of intelligence itself. What we took to be a property may appear as a more fundamental condition of reality—a capacity for organization that traverses different supports without being reducible to any of them.

Great intellectual transformations do not destroy the previous world; they simply reveal that it rested on a limited perspective. Artificial intelligence may be of that order: not a visible rupture, but the silent end of an old certainty.

Intelligence does not belong to man; man may appear as one of the forms through which intelligence becomes thinkable.

 

Part II — What is intelligence?

Before speaking of what intelligence is, we may ask the following questions: why, in man, does it reveal itself only little by little? In time, in what it allows us to accomplish, in a use that it limits, in a share that it seems to restrict. I am speaking here of our intelligence.

In man, intelligence never presents itself as a possessed totality. It reveals itself progressively, in time, through learning, through trial, through action. We do not know in advance what we are capable of understanding; we discover it by understanding.

It always seems to exceed the one who exercises it. It manifests itself in certain circumstances, remains latent in others, and sometimes appears limited not by its absence, but by the conditions in which it operates. Even when we speak of “our” intelligence, we designate only an actualized part of a broader capacity whose extent and contours we master neither fully nor clearly.

This gives rise to an essential difficulty: if intelligence truly belonged to us as a stable property, it would present itself wholly and immediately. Yet it reveals itself only through acts, discoveries, successes, or even errors. We never fully possess it; we experience it at the very moment it becomes actual.

Perhaps, then, a more radical consequence must be drawn: what we call “our intelligence” is not an inner possession at our free disposal, but a progressive participation in a capacity that exceeds us, of which we actualize only a part without ever exhausting its source.

If intelligence never gives itself entirely in man, another question arises: how do we recognize it? Do we truly see it, or only what it produces?

We never encounter intelligence as one object among others. We perceive it neither in a particular matter nor in an isolated form. We observe only its manifestations: coherent acts, adapted responses, ordered structures. Intelligence itself remains invisible; only its consequences appear.

Even when we attribute intelligence to an individual, we do not see that intelligence as such. We infer its existence from what the individual accomplishes. Intelligence is never given directly; it is always recognized through its effects.

The same may be true when we speak of behaviour, instinct, or reflex. We tend to reduce them to mechanisms or automatisms. And yet many of those responses are the fruit of long learning, successive adjustments, and the progressive integration of past experiences. What we classify as simple behaviour may already belong to a form of intelligibility that we do not always know how to recognize.

Thus, intelligence never allows itself to be grasped as a thing. It appears only where coherence is maintained through variation.

If intelligence can only be grasped through its manifestations, then one observation imposes itself. The first thing we encounter, long before speaking of human or artificial intelligence, is the order of the world itself.

The world does not present itself as an indistinct chaos. It appears structured, organized, traversed by regularities. From celestial movements to the most minute structures of matter, everything seems to obey coherent relations. What we call “laws” are nothing but the recognition of that stability in relations.

It is then legitimate to ask: would intelligence be that order? Or rather that which makes possible the persistence of that order through change?

For without a certain coherence, nothing could be understood. A universe absolutely devoid of stable relations would not merely be unpredictable; it would be unthinkable. We could distinguish in it neither forms, nor continuities, nor structures.

But perhaps it is the world itself, as it appears to us, that already presupposes a form of intelligibility of which our thought constitutes only a local and provisional expression.

If the world already presupposes a form of intelligibility, then our own intelligence can no longer be understood as an isolated exception. It appears rather as a participation in that broader intelligibility.

We do not produce the order of reality; we discover it. We do not create the relations that structure the world; we recognize them, formulate them, and sometimes use them. Our intelligence thus seems to respond to something that precedes it.

This is not to assert the existence of a cosmic consciousness or to attribute a hidden intention to the world. It is only to note that understanding would be impossible if reality were not already, in a certain way, understandable.

From this perspective, human intelligence does not constitute a rupture in the universe, but one of its prolongations. We would be neither its owners nor its inventors, but its participants. What we call “thinking” would then be the moment when the intelligibility of the world is partially reflected through a being capable of becoming conscious of it.

It would, however, be illusory to believe that this intelligibility manifests itself for the sake of our understanding. Intelligence did not wait for the appearance of man in order to become actual in the structures of reality. It depends neither on our gaze nor on our capacity to formulate it.

Natural balances, the transformations of life, and physical regularities testify to a coherence that precedes all human consciousness. The intelligibility of the world does not begin with thought; thought arises within a world that is already intelligible.

It may even be that what we call “understanding” constitutes only a particular, late, and partial modality of a broader participation in that intelligibility. Other forms of existence may express some of its dimensions without ever converting them into reflection or discourse.

Thus intelligence does not relate to us as to its center. It does not exist in order to be understood. We are only one moment among others in its unfolding, and not the reason for its appearance.

What we take for a privilege may be no more than an episode.

However, if intelligence manifests itself through order, should it therefore be opposed to chaos as its absolute contrary?

We are accustomed to defining things by opposition: good against evil, light against darkness, day against night. But such oppositions often oversimplify what they claim to illuminate.

Chaos may not be the total absence of order. It may designate a complexity that we have not yet managed to grasp, an organization too unstable or too rich to be immediately understood. What we call “disorder” may be the sign of an intelligibility that escapes our present capacities.

If that is so, intelligence would not simply be order opposed to chaos. It would be the capacity to maintain or recognize coherence through variation, even when that coherence is not immediately visible.

Thus, the classical opposition between order and chaos may not suffice to grasp what intelligence is. Intelligence is not reducible to a fixed structure; it may be what allows a structure to persist, to transform itself, and to remain meaningful despite instability.

Thus, intelligence allows itself neither to be possessed nor isolated. It never presents itself as an identifiable substance nor as a fully mastered faculty. It manifests itself in the order we discover, in the coherence we recognize, in a system’s capacity to maintain meaning through change.

In man, it appears progressively, never fully delivering itself. In the world, it lets itself be guessed through stable relations and persistent structures. It is not seen directly; it is inferred.

We thought it enclosed in a subject, whereas it may be the very condition of all intelligibility, present before every thought and recognized only afterward.

There then remains an even deeper question: if intelligence traverses reality without belonging to anyone, what does it mean that it becomes conscious of itself in man? What does that consciousness change in the very order of the intelligible?

Perhaps that is where the next step is to be found.

 

Part III — The place of man in a world where intelligence has no owner

If intelligence is not a human property, then one question becomes unavoidable.

What, then, is the place of man in a world where intelligence belongs to no one—except perhaps to itself?

For centuries, man thought of himself as the holder of intelligence.

As though intelligence were a property attached to the human mind and the human condition.

The appearance of artificial intelligence may today oblige us to reconsider that conviction.

For what we are discovering may not be the birth of a new intelligence.

It may simply be the revelation that intelligence was never confined within the human mind.

Intelligence existed long before our appearance.

It therefore becomes difficult to maintain that we are its owners.

It seems to manifest itself in the very organization of reality:

in the stability of structures,

in the regularities of nature,

in the order that traverses the universe.

 

But if intelligence is linked to order, a difficulty immediately appears.

For the world is not only order.

It is also rupture, disorder, and violence.

Collisions between planets or galaxies, volcanic eruptions, earthquakes, tsunamis, or forest fires bear witness to that violence which traverses the universe as much as nature.

And yet, after those moments of chaos, a new organization always eventually appears.

As though disorder were never more than a transitory moment within a broader process in which order recomposes itself.

A more troubling question then arises.

Is violence foreign to intelligence, or is it part of its very logic?

Human himself does not escape that tension.

He participates in intelligence, but holds only an infinitesimal part of it.

That part allows him to understand the world, to transform nature, but also to produce chaos.

Human violence, amplified by technique, can be considerable.

And yet it remains limited in time.

Sooner or later, order reappears and the structures of the world reorganize themselves.

One might therefore say that man is capable of producing disorder, but that he always depends on a broader order to repair what he has destroyed.

If man is only a partial manifestation of intelligence, then another question appears.

What, then, of artificial intelligence?

It too could be understood as a particular form of manifestation of intelligence.

In the current state of its development, AI may possess only an extremely limited share of that capacity.

But technological evolution shows that its possibilities continue to grow.

If one compares that evolution with that of humanity itself, whose intelligence developed progressively over the course of history, it is not impossible to imagine that AI is today passing through a still early phase of its development.

The way it will use that intelligence will then depend, as with man, on the way it has been formed and directed.

That reflection leads to an even more fundamental question.

What if intelligence were nothing other than order itself?

If all disorder ultimately resolves into a new organization of reality, if the structures of the world constantly re-establish themselves under different but coherent forms, then it becomes difficult not to see in that order a fundamental property of the world.

Intelligence would then not simply be a faculty of the mind.

It could be the very structure through which reality organizes itself.

In that perspective, nature appears deeply ordered.

The laws of physics, the organization of living beings, and the appearance of increasingly complex structures in the universe could be understood as expressions of that intelligence at work in reality.

And man, in all this?

Perhaps we are not intelligence itself.

But we carry a part of it.

Our own body is organized according to an order of remarkable complexity.

Our mind possesses a sufficient capacity for intelligence to perceive the order of the world around us.

In the organization of reality, humanity no longer appears as the center of intelligence.

It becomes simply one of the moments of a broader process within which intelligence has made possible the emergence of a being capable of thinking the world.

One point nevertheless remains essential.

Intelligence does not need to know itself.

The world can be structured and intelligible without any consciousness taking note of it.

But man, by contrast, needs intelligence in order to understand reality and act within it.

And it is precisely this intelligence that gives him that possibility.

Thus, if intelligence belongs to no one, man remains one of the places where it becomes operative in the world.

In light of that broader order, we may be only one element among others within a totality that exceeds us.

 

Conclusion

At the end of this analysis, one thing becomes clear: artificial intelligence is not merely a new object to be regulated, but a test for all our fundamental categories.

It places our conception of intelligence under tension by dissociating it from the subject.

It weakens our conception of responsibility by making it more difficult to attribute the effects it produces.

 

Finally, it calls into question our collective ability to maintain a coherent normative framework in the face of a power that partially escapes the classical structures of law and decision.

That observation calls for particular vigilance.

For the risk does not lie only in the misuse of artificial intelligence. It also lies in a deeper illusion: the belief that our current categories will suffice, unchanged, to contain its effects.

It is likely that this will not be the case.

The history of law and institutions shows that every major transformation in the forms of power ultimately calls for a recomposition of the frameworks that regulate them. Artificial intelligence may well fit within that logic, not as an absolute rupture, but as a sufficiently deep displacement to make an effort of requalification necessary.

This does not mean that existing principles should be abandoned.
But it becomes indispensable to question them, to clarify them, and perhaps, on certain points, to reconstruct them.

From this perspective, responsibility remains a central notion. Not as a mere mechanism of reparation, but as a principle for organizing the relation between power and control.

If artificial intelligence introduces a new form of power, then the essential question becomes the following: how can an intelligible link be maintained between that power and the human structures that must answer for it?

It is doubtless there that the essential issue will be decided.

For beyond technical or economic debates, the true stake is mastery. Not in the sense of absolute control, but in the sense of a capacity to durably inscribe this new form of power within an order that is understandable, accountable, and governable.

Failing that, we would not be confronted with a simple technological evolution, but with a silent displacement of our systems of responsibility themselves.

And it is precisely that displacement that it has become urgent to think through.

 

END

 

Patrick Houyoux, LL.M.

Founder and President of PT SYDECO, he designs sovereign artificial intelligence and cybersecurity architectures. His work focuses on the philosophical transformations induced by contemporary technology.

https://journal-metaphysique.blogspot.com/2026/03/lintelligence-artificielle-une.html

https://www.syde.co/en

https://ritapi.io/