Should a recognized work cease to change?

The question might seem absurd. After March 20th and then December 18th, 2025, it became one of the most important questions in the entire history of the Abode of Chaos.

For over twenty years, we had fought against a very identifiable threat: that of having to go back. The restoration process carried within it the idea that there must be a previous state of the place to which we would have to return. For me, this operation would have meant destroying a constituent part of the work in order to rediscover what existed before it. Then the situation was reversed.

The artistic value of the Abode of Chaos is recognized. Its status as a Total Work of Art is acknowledged. Its evolving nature is itself part of the vocabulary by which it is now viewed. And suddenly, a danger appears, almost symmetrical to the previous one.

Yesterday, one might have wanted to return to the state before the work. Tomorrow, one might want to return eternally to the state of the day the work was recognized.

In both cases, the problem would be the same. Time would stop.

Now, to mutate also means to survive what seeks to protect it. The paradox is immense.

But the Abode of Chaos is made of time. Not just because it ages. All materials age. Stone erodes. Metal oxidizes. Wood warps. Paint changes. Concrete transforms. But the Abode of Chaos possesses a more radical temporality: its evolution belongs to its artistic process.

Since 1999, it has absorbed the world: wars, revolutions, catastrophes, economic and technological changes, the faces that make History, the dead and the survivors, the powerful and the damned, data, networks, drones, and then artificial intelligence.

Therefore, it does not simply represent one era. It is constructed in the transition from one era to another.

If we were to stop this process, we might preserve an extremely rich material state. But we would eliminate one of the constitutive principles of the work.

That is why recognition does not resolve the heritage issue. It finally makes it possible.

Before recognition, the priority was survival.

After recognition begins the much more difficult question of transmission.

This is precisely what the researchers who worked on the Abode of Chaos understood before this question became institutionally unavoidable.

Claudine Houbart and the DIVA group from the University of Liège have brought a perspective here that deeply interests me because it rejects the easy solution.

The Abode of Chaos is not approached as an object that can simply be stabilized and then preserved. It is viewed as a complex object: living, evolving, unique, and unreproducible. This non-reproducibility here changes the nature of the heritage issue. A transportable sculpture can change location while retaining a crucial part of its identity.

The Abode of Chaos cannot be moved. You cannot dismantle nine thousand square meters and rebuild them elsewhere like a stage set. Even with unlimited technical resources, physical reproduction would not suffice. Because the artwork has a relationship to its location.

To its history. To its buildings. To Saint-Romain-au-Mont-d’Or. To the topography. To the successive layers. To the transformations. To what has been destroyed. To what has been added. To what has aged. To what has oxidized. To what has been repainted. To what became almost invisible and then was made visible again.

We could reproduce a form. We could not reproduce its time.

That is why the notion of irreplaceability has a precise meaning here. If the Abode of Chaos were to disappear materially, photographs, films, plans, surveys, point clouds, testimonies, archives, and billions of data points could still be preserved.

Tomorrow, we may be able to produce a simulation of extraordinary precision. But we would no longer have the Abode of Chaos. We would have its memory.

In the digital age, we must maintain the distinction between recording a thing and preserving its existence.

We had built the digital twin so as not to lose the successive states of the Abode of Chaos. It could preserve the trace of a surface before its transformation, of a work before its relocation, of a state of the place before another came to cover it. But it did not preserve the work in its place.

It preserved the ability to compare what changes.

This difference proves decisive when thinking about transmission: documenting a state does not mean declaring it definitive.

Imagine a statement taken at a given date. Then another one taken a few years later. The two statements can be compared.

A sculpture has changed. A surface has been transformed. A new intervention has appeared. A structure has aged. A color has changed. A work has been added. The transformation itself becomes documentable.

We are therefore no longer forced to choose between two extremes: freezing the work to preserve its memory, or letting it evolve at the risk of losing knowledge of what it was. We can begin to imagine a third way.

Preserve the movement and document the states.

This idea seems crucial to me for 21st-century heritage. It goes far beyond the Abode of Chaos. Our societies are producing more and more works that no longer correspond to the traditional model of the stable object.

Installations, digital works, software and environments.

Participatory works, generative creations and algorithmic systems.

Works that depend on a network, evolve with data, or whose behavior changes.

What does it mean to preserve these objects?

Save their code?

Their equipment?

Their behavior?

Their execution environment?

Their documentation?

The artist's intention?

The interactions they produced?

The problem facing the Abode of Chaos is therefore part of a much broader transformation of heritage. Heritage has long been about preserving objects. It will increasingly have to learn to preserve processes.

This is where the concept of living heritage takes on its full meaning. It must not become a seductive but empty phrase. Living heritage is not simply heritage where people continue to move about. It is alive when certain transformations become part of its identity.

But this definition immediately raises a problem.

If all change is permitted in the name of life, how can we distinguish evolution from destruction?

That is the real question. And I do not want to avoid it. Saying that the Abode of Chaos is evolving does not mean that just anyone could do whatever they want there tomorrow.

Evolution is not arbitrary. Mutation has a genealogy.

A living work is not a work without integrity. On the contrary, it possesses an integrity that is more difficult to define.

In a fixed work, integrity can be largely material.

In an evolving work, it also becomes conceptual.

We must therefore return to the cosa mentale. From the beginning, the Abode of Chaos has possessed an architecture of thought. The thousands of works do not constitute a random collection. They belong to a system.

A semiology and a semantics linked to scientific Chaos, alchemy, contemporary History, the aesthetics of catastrophe, vanities, geopolitics, memory, information, networks and mutation.

This system allows us to distinguish evolution from simple modification.

The shift to pink has already demonstrated this: a profound transformation can change the perspective without breaking the grammar of the work. The challenge, therefore, is not to prohibit change, but to determine whether it stems from the conceptual continuity of the Total Work of Art.

But this logic raises a question that concerns me personally. What will happen when I am no longer here?

This is probably the most difficult question in this chapter.

As long as I am alive, I can say what is part of my cosa mentale. I can decide, refuse, correct, go back, experiment.

But a work that claims a place in the long span of time must think through the moment when its author disappears.

Otherwise, it is not truly thinking through its transmission.

I reject two solutions. The first would be to say: after me, everything is permitted. That would be to abandon the work. The second would be to say: after me, nothing should change. That would be to kill it.

We must therefore build a third way. It is not yet fully written. And I believe it should not be written too soon. But its principles are beginning to emerge.

Documenting the cosa mentale, the materials, the techniques, the works, their positions and their transformations. Preserving the archives, the digital models, the photographs, the films, the texts, the testimonies, the specific recipes and the protocols. Above all, preserving the logic that allows us to understand why an intervention belongs or does not belong to the work.

This requires much more than a will. It requires a structure for inheritance planning. And part of this structure already exists.

Nadège is a living memory of it. She has known the Abode of Chaos since its inception. She has witnessed the decisions. She knows the laboratories. The materials. The trials. The errors. The setbacks. She knows how much an intervention that seems spontaneous can conceal months or years of work.

Sydney has a different relationship to the work. He grew up within it. He constructs its mediation. He belongs to the generation that will have to receive this memory without simply repeating it.

Teams possess their knowledge. Individuals possess theirs. Archives hold what no individual memory can preserve.

The digital twin possesses spatial memory. Films possess movement. Photographs possess states. Texts possess formulated intentions. Interviews possess orality. And now this book itself enters into this architecture.

Dialogue Between a Thinker and AI is not the story of a life. It is a world-book, the trace of a thought in the process of forming, reflecting upon itself, correcting itself, and confronting a non-human intelligence. The book thus becomes one of the means of transmitting this experience: not preserving an existence, but preserving the movement of a thought.

I speak. I come back. I correct. I clarify.

I reject an attribution when it is false. I distinguish what I thought from what Nadège developed. I distinguish what Sydney interprets from what he created. I search for the documents. I refuse to quote a letter until I have found it.

This obsession with accuracy is not merely editorial. It is about preserving heritage. We are building a memory that must outlive those who produce it.

And this is where artificial intelligence begins to appear differently.

Not yet as a subject. Not yet as an exhibition. But as a problem of memory.

What happens when thousands of documents, images, plans, videos, testimonies, surveys, and data can be queried not only through keywords, but through relationships?

Throughout my professional life, I have built databases. I have organized corpora. I have designed tree structures. I have thought of information as a material.

Artificial intelligence adds a new dimension. It can begin to explore the relationships among these materials on a scale that the human mind alone cannot exhaustively explore. That does not make it the author of memory; it can become an instrument for traversing it.

An archive that no one knows how to query is a dead memory. A searchable archive becomes a space of knowledge. I have dedicated my life to this transformation.

The Judicial Server was already doing this. Artprice does this on a different scale. The digital twin does this in space. And artificial intelligence now opens up the possibility of doing it in dimensions we could previously only glimpse.

I see a strange architecture emerging. The physical Abode of Chaos preserves matter. The digital twin preserves space. The archives preserve states and evidence. Testimonies preserve perspectives. Texts preserve concepts. And artificial intelligence can begin to build passages between these memories.

It does not replace anything. It connects.

Perhaps this is where part of the answer to the heritage paradox lies.

To preserve a work that must never end, we must stop believing that preservation means choosing a definitive image. We must build a multidimensional memory capable of accompanying its transformation.

Preservation then functions almost like a documentary nervous system.

It records. Compares. Alerts. Recalls. Reconnects.

It allows the present to know what the past was without preventing it from becoming something else. This is exactly what an evolving work needs.

The 2025 recognition then takes on a much more demanding meaning for me. It does not say:

"Now, let's protect what exists and leave everything else alone."

It forces us to ask a much more difficult question:

How can we guarantee that what exists today will still be meaningful tomorrow without losing the memory of what it once was?

This question concerns the Abode of Chaos. It also concerns our societies.

We are obsessed with preserving data and extraordinarily bad at preserving its meaning.

We store everything: billions of photographs and messages, videos, documents, copies, backups.

But accumulation is not memory. Memory presupposes a structure. A hierarchy. Relationships. An ability to retrieve. An ability to understand the context. Above all, an ability to distinguish the trace from its interpretation.

This is why the archives of the Abode of Chaos must remain linked to their genealogy. An undated photograph can be misleading. A work without context can change its meaning. A quotation without a source can become a legend. An approximate attribution can, after twenty years of repetition, become a false truth.

We have just experienced this with the color pink. If we had allowed a simplified attribution of its origin to take hold, a few decades might have been enough to transform this simplification into the official narrative.

Corrections were needed. The exact wording was required. The cosa mentale. The impulse. The research. The chromatic and technical developments.

Successive contributions. Mediations. External perspectives. Contemporary documents.

Gilles Blanckaert's message, found and dated July 2022, now offers us an almost perfect example.

This great observer had immediately perceived what we would later take years to formulate in its full scope: pink as a revealer, as a break from uniformity, as a possibility of making the space of the Abode of Chaos stand out again.

The document existed. But until it was found, I refused to reconstruct its words from memory. Now that it is here, it can take its rightful place in history.

That is what passing on knowledge means to me.

It is not just about preserving a lot. It is about preserving just enough.

A memory does not become reliable simply because it contains a large number of elements. It becomes reliable when it allows us to distinguish between different levels.

This relates to the fact, the author's words, or the interpretation.

This relates to contemporary testimony, retrospective reconstruction, or hypothesis.

This relates to measurement, the work, or its history.

After those who experienced the events have disappeared, documents will begin to speak in their place. The precision of attributions, dates, and contexts therefore underpins reliability.

And a misattributed document can unintentionally lie. An approximate date can shift the causal link. A photograph taken out of context can fabricate a story. A testimony paraphrased for thirty years may end up bearing no resemblance to what was originally said.

Artificial intelligence makes this question even more dizzying.

Because it dramatically increases our ability to find, connect, and reformulate. But the power of a system to connect information does not guarantee the truth of what it connects.

If the source is false, it can industrialize the error.

If the attribution is wrong, it can be repeated on a huge scale.

If the context is gone, it can produce a seductive yet historically false relationship.

The heritage problem therefore also takes on an epistemological dimension. How do we build a memory rich enough to remain alive, yet rigorous enough not to manufacture its own mythology?

This question ties into my entire reflection on bias and the blind spot. Archives, too, have their blind spots. They preserve what someone has decided to preserve. They name what someone has decided to name. They classify according to a structure. They therefore already possess a worldview.

We need to be aware of this. Technology does not eliminate this problem. It makes it visible on a different scale. I recognize here something that has been a constant throughout my life.

Rankings are never neutral.

A tree structure produces a way of thinking. A thesaurus produces a way of connecting. A database produces a way of retrieving. An algorithm produces a way of prioritizing. And now, an artificial intelligence model produces a way of traversing spaces of knowledge.

The question, then, is:

What kind of memory architecture do we want to convey with the Abode of Chaos?

I do not want a hagiography. I do not want a documentary mausoleum. Nor do I want a future intelligence to reconstruct the Abode of Chaos solely from the legend of its author.

I want it to be able to encounter the contradictions. The trials. The errors. The transformations. The documents.

External perspectives. Conflicts. Decisions. Successive versions.

Matter as it was. Thought as it was formulated. And above all, the relationships between these elements.

Because that is where life is.

A living work is not just a material that continues to change.

It is a work whose very memory remains capable of producing new readings without losing the trace of what it was.

This idea leads me to a conclusion that would have been unthinkable at the beginning of the Abode of Chaos.

Conservation may no longer be just a function that occurs after creation.

It can become part of the creation.

Documenting can become an artistic act when it allows an evolving work to preserve the memory of its metamorphoses.

Building the digital twin is not making a copy of the Abode of Chaos. It is building one of the conditions of its transmission.

Archiving photographs is not about accumulating images. It is about preserving their states.

Recording lyrics is not about creating a legend. It is about preserving the conditions under which a thought was formulated, and sometimes corrected.

And bringing artificial intelligence into this system is not about handing it the keys to the work. It is about creating a new tool for navigating its memory.

That is why I no longer see the issue of heritage as an opposition between conservation and creation. At the Abode of Chaos, they will have to learn to become complementary.

Keep it so you can continue.

Continue without deleting.

To change without forgetting.

To transmit without fossilizing.

Document without reducing.

And to prepare for the moment when the work must be able to produce meaning beyond the physical presence of the one who conceived it. This is an immense responsibility. But it also contains an extraordinary possibility.

If we succeed, the Abode of Chaos will not only be one of the works that posed the question of preserving a living process to 21st-century heritage. It could become a laboratory for this question. A physical and documentary laboratory. A digital and memorial laboratory. And now, an algorithmic laboratory.

For at the very moment the State recognizes the Abode of Chaos as a “Total Work of Art, in perpetual evolution,” a new transformation is already underway. The question of its transmission beyond the presence of its sole creator remains to be addressed: I do not wish to confine the ministerial letter to that question. Artificial intelligence is not waiting at the Abode of Chaos's door. It has already entered. And with it arises a new question.

Not only that: how to keep a work alive?

But what becomes of a living work when a new form of intelligence enters its own system of memory, creation, and questioning?

The answer cannot be theoretical. It must be tested. The Abode of Chaos will do it.

The work begins again. From Chaos, AI Is Born.