How an age restriction in a competition led to the emergence of Noah33, why artificial intelligence became not an autonomous curator but part of a distributed system, and how this experiment reshaped the understanding of the curatorial profession.

In the EDEN.exe project, presented in Berlin at the 48 Stunden Neukölln festival, artificial intelligence did not select artists or make final decisions. Curator Yulia Sysalova and her team created Noah33 - an artistic and conceptual system through which they explored the limits of machine interpretation, distributed authorship, and human responsibility. In an interview with Art MUSE, Julia discusses how EDEN evolved into EDEN.exe, why the exhibition became a “post-curatorial fiction,” and what happens to the curator’s profession when an algorithm becomes involved in the production of meaning.

More about “EDEN” project

To learn more about the exhibition and its works, you can explore via the link

What personal or professional question did the EDEN.exe project begin with for you? What did you want to explore in artificial intelligence, the curatorial profession, and the relationship between humans and technology?


First and foremost, EDEN.exe is an experiment and, in a certain sense, a speculative scenario - a hypothesis I proposed to artists and audiences to explore with critical distance and a degree of self-irony.

I am interested not in technological innovations as such, but in cultural transitions - those moments when society begins to develop new ways of producing memory, identity, and collective imagination. Art allows us to perceive these shifts before they receive philosophical or political articulation. That is why we did not use artificial intelligence literally as an autonomous curator who assigns tasks to artists or makes decisions on behalf of a human team. Noah33 became an artistically reimagined form of dialogue with AI, enabling us to explore the meanings, limitations, and contradictions of such interaction.

The idea for the project emerged about a year before its realization in Berlin. The first stage was CODE ERROR, created for The Wrong Biennale. Even then, I was interested in the notion of error as an event that disrupts the stability of a system and provokes its evolution. The project was selected for the biennale, presented in Miami and Los Angeles, and is currently a finalist of the OSTEN Biennale 2026 in Skopje.

Today, we are once again experiencing a cultural transition: for the first time, computational systems are participating alongside humans in the production of images and meaning. It is precisely this shift that became the subject of EDEN.exe.

What came first: the idea of involving AI in curatorial work, the theme of the exhibition, or a specific circle of artists and artworks? How did the initial concept evolve during the preparation process?

The theme came first - linked to an exploration of how humans construct mechanisms of avoidance, utopian models, and ideas of ideal space. We tested the first version of the project, Eden. Lost Coordinates, in Athens and realized that it held potential for further development.
The circle of artists and some of the works already existed prior to the project in its current form. At the same time, AI was not added to a finished exhibition as a spectacular technological element. It became the next step in the research: if paradise is a system created by human imagination, can an algorithm offer us its own version of such a system? Can artificial intelligence become a participant in shaping a new collective imagination?

When the opportunity arose to take part in the competition of the 48 Stunden Neukölln festival, I realized that this was precisely the context in which EDEN could be presented in a more experimental and contradictory form. For me, an exhibition is a space where an idea is tested, and the festival environment allows for the exploration of hypotheses that are not required to become final statements. This is how EDEN transformed into EDEN.exe: from a reflection on the formation of human utopias into an experiment involving the inclusion of non-human intelligence in the process of dialogue. In this form, the project received a festival grant for its realization.

How did the title EDEN.exe and the definition “post-curatorial fiction” emerge?

The emergence of this form of the project was influenced by a specific limitation in the original competition conditions: the application had to be submitted by a curator under the age of 35. I did not meet this age criterion, but the artists wanted to try submitting the project. During our collective discussions, the initial idea arose to create a co-curator avatar as an independent artistic and curatorial persona. This is how Noah33 came into being.
The situation itself became the starting point of a curatorial fiction. However, quite quickly, this fiction ceased to be merely a response to an institutional limitation and raised a more significant question: can this emerging form of collective consciousness become a response to the demands of our time?

The title EDEN.exe brings together two different systems. Eden is one of the oldest archetypes of human thought, embodying the idea of an ideal world order. The .exe extension, on the other hand, refers to an executable computer file—a program that must be launched in order to begin operating. We see the project not as a finished model, but as an evolving structure.
The definition “post-curatorial fiction” emerged because the project blurs the familiar figure of a single authorial voice. Here, meaning is produced through an exchange between the meta-curator, the AI curator, the artists, the artworks, the space, and the viewers. “Post-curatorial” does not imply a rejection of professional or human responsibility. Rather, it describes a situation in which the curator is no longer the sole center of meaning production, and curating becomes a distributed system of negotiation.

Why did you define your own role as that of a meta-curator?

The prefix “meta” derives from the Greek μετά and allows for several directions of meaning: “between,” “after,” and “beyond.”
On the one hand, the meta-curator exists within the process—between artists, artworks, physical space, and digital reality. On the other hand, they must be able to step outside this system and observe how connections are formed within it, how roles are distributed, and how decisions are made. Therefore, my task was not limited to creating the concept and the exhibition. I observed the process itself: how humans interact with AI, where technology expands human thinking, and where it creates an illusion of autonomy. In EDEN.exe, the meta-curator is responsible both for the final statement and for the critical framework of the experiment.

How was the project team formed, and how were roles distributed between you, curatorial assistant Veera Romanoff, Noah33, and other participants? Who was responsible for which decisions?

The project grew out of an ongoing dialogue with artists within the framework of a previous exhibition, and we did not search for participants specifically for the application. It is important for me to discuss the concept with the participants: they are not executors of a pre-written curatorial script but active contributors to its development.

My role consisted of developing the overall conceptual framework, coordinating the physical and digital layers of the project, and maintaining a critical distance in relation to AI. Veera Romanoff was responsible for the spatial composition and, together with me, translated the concept into the physical dramaturgy of the exhibition, which was installed in the building of the former shopping center CANK (Kulturnetzwerk Neukölln e.V.). Noah33 represented the “non-human” layer of this system and participated in the interpretation of selected works. The artists, in turn, were responsible for the development and physical realization of their own pieces.

Thus, the project was structured not as a vertical hierarchy with a single center of control, but as a distributed system in which each participant had their own area of responsibility. At the same time, final curatorial responsibility remained with the human team.

Did the artists work with pre-existing works, or did they create and adapt them specifically for EDEN.exe? What did they know about the role of AI, and were they able to challenge its interpretations and decisions?

EDEN.exe included both works from the previous iteration of the project and pieces created or adapted specifically for the new exhibition. All the artists were aware of the role of AI and were able not only to challenge its interpretations but also to reject them. Each participant had their own research question and their own mode of engaging with the technology.

What was Noah33 technically? Was it an artistic persona, an existing language model, a specially configured AI agent, or a system composed of multiple tools? Who participated in its creation and configuration?

In its final version, Noah33 was not a separate autonomous product, but rather an artistic and conceptual system that existed through the relationships between people, artworks, and digital tools. The name itself refers to Noah and the Ark—a structure designed to carry humanity through catastrophe. In the project, we ask whether AI, a collective consciousness created by humans, can preserve the possibility of a future. In a certain sense, the entire exhibition itself was Noah33.

At the initial stage, we thought of Noah33 as a specific digital avatar. The idea was that it would have a visual appearance, a voice, and the ability to communicate with viewers. However, during the process, I realized that by giving AI a human appearance and human-like behavior, we inevitably turn it into a simulation of a person. This would have significantly simplified the question we wanted to explore, so we abandoned the idea of a literal anthropomorphic avatar.

The number 33 is connected to the history of technology: a classic long-playing vinyl record is played at a speed of 33⅓ revolutions per minute. Therefore, 33 became for us a symbol of mechanical reproduction, repetition, and technologically organized memory. Although it may also evoke other symbolic associations.

In a certain sense, the entire exhibition itself was Noah33.

Who formulated the instructions for interacting with Noah33, and how was AI used at different levels of the project?

The instructions for the dialogue were developed by us. At the level of individual artistic practices, AI was used for conversations, working with digital archives, restoring materials, or engaging with online data. At the curatorial level, it helped reveal the limitations of algorithmic intelligence and made visible the very process of interaction between humans and the technological systems they create - including our expectations of these systems, our projections, fears, and our tendency to attribute agency to a machine.

What did your interaction with it look like in practice? Was it a sequence of dialogues and refinements, or did the system generate solutions that the team then accepted or edited?

It was a sequence of dialogues, refinements, and evaluations. Artificial intelligence did not provide us with a ready-made exhibition that the team would simply accept or edit. We established the framework, formulated questions, received responses, and decided which of the proposed connections could genuinely function within the project. At the level of individual artworks, the interaction could be more direct, while many works did not involve AI technologies at the level of their form at all.

Could you give two examples of proposals that emerged in dialogue with Noah33: one that entered the exhibition almost unchanged, and one that the team rejected or significantly reworked?

One example of a solution that entered the project almost unchanged was the system of labeling the artworks: N01–N32 + N33, where the exhibition itself was understood as the thirty-third work. This formula reminded viewers that they were entering a unified system rather than simply visiting an exhibition of separate objects.

At the same time, I cannot point to a single rejected proposal, because more often we rejected a particular type of interpretation: one that was formally convincing but too smooth and universal. Such texts could be applied to almost any artwork and therefore did not withstand the curatorial examination.

What errors, limitations, or unexpected interpretations emerged during its work? Was there a case when AI noticed a connection in an artwork that the human team had not seen?

Yes, there were such cases, and they were precisely the ones we found the most interesting. For example, in the work of Nelya Akimova, which reinterprets the image of the Venus of Willendorf. For her, this figure was primarily an image of a kind of cocoon within which a new form of humanity is being formed. In the dialogue with AI, another interpretation emerged: the network surrounding Venus was read not as a protective structure, but as a system of restrictions within which the very source of life becomes enclosed.

This interpretation did not alter the artwork itself, but it allowed us to see how differently human and digital forms of intelligence are capable of constructing networks of meaning. At the same time, this was also where the main limitation of AI became apparent—it tends toward complete explanations and struggles to preserve the productive ambiguity upon which contemporary art is often based.

Another case was connected with the image of Noah33. Artificial intelligence proposed the figure of Noah as a symbol of humanity’s salvation, but this was precisely where the fundamental irony of the project emerged. We do not claim that technology can become a new ark. On the contrary, the exhibition raises the question of what can truly become a form of salvation today: further technological development, a critical dialogue with technology, or the human ability to preserve freedom of choice?

How was the concept of EDEN.exe translated into physical space? Who determined the dramaturgy of the exhibition, the relationships between the works, the viewer’s route, the lighting, and the installation process? And which artwork or combination of works became its conceptual center?

The space of the former shopping complex played an important role in this process. Having undergone a transformation of its function, it itself became a metaphor for the cultural transition from previous models of collective life to new temporary forms of existence.
It was important for us that the viewer experienced the exhibition not as a sequence of separate artworks, but as a unified environment. We designed not a route between objects, but a gradual transformation of perception according to the formula N01–N32 + N33. Moving through the space, the viewer encountered works in which it became impossible to draw a clear boundary between memory and data, between the material object and the digital image.

In a certain sense, the central artwork of EDEN.exe was the system of relationships between the works themselves. Individual pieces retained their autonomy, but only within the exhibition space did they begin to produce new meanings that could not be discovered when each work was viewed in isolation.

Did the final exhibition function as an independent artistic statement for viewers who knew nothing about Noah33? Or was understanding the mechanics of AI-based curating an essential part of experiencing the project?

Yes, the exhibition was intended to function as an independent artistic statement. Knowledge of the entire mechanism of interaction with Noah33 expanded the viewer’s perception, but was not a necessary condition.

We stated that we were exploring errors, failures, and instability within human-created systems. However, not all works were directly connected with AI or digital technologies. The notion of error could manifest itself in memory, in the perception of the past, in humanity’s attempt to create an ideal order, in relationships with nature, or in the very model of utopia. Therefore, viewers could enter the project through a particular artwork, material, image, or personal association, and only later discover the broader framework of dialogue between physical and digital reality.

Was there a risk that technological novelty and the figure of the AI curator would overshadow the artists and the artworks themselves? How did you try to maintain this balance, and how did the participants respond to it?

Such a risk exists if artificial intelligence is given the authority to think instead of the artist or to create works on their behalf. In that case, technology quickly becomes the main attraction, while the artistic statement turns into merely an illustration of its capabilities. This did not happen in EDEN.exe, because AI was not the final decision-making authority. We tried to keep the focus not on the claim that artificial intelligence created the exhibition, but on how its presence changes our understanding of authorship, interpretation, and responsibility.

How did viewers learn which decisions belonged to Noah33 and which were made by the human team, and which reactions were the most revealing for you? By what criteria do you evaluate the result of EDEN.exe—as an exhibition and as a curatorial experiment?

We did not aim to label every decision: this was made by a human, and this was suggested by AI. For us, it was more important to place the viewer in a state of technological liminality between physical and digital reality, where the boundaries of authorship are not always obvious. At the same time, all accompanying materials clearly stated that the project was built as a dialogue between human and artificial intelligence. The viewer was aware of the conditions of the experiment but was invited to observe independently how these relationships manifested themselves within the exhibition.

We evaluate the result primarily through the audience’s response. EDEN.exe became one of the notable projects of the festival. Viewers spent time in the space, carefully read the texts accompanying the works, asked questions, returned to individual pieces, and entered into dialogue with the team. For us, it was important not simply to attract attention through the idea of an AI curator, but to sustain that attention through the artworks themselves and the meanings that emerged between them.
The second criterion is research-based. The project allowed us to understand which ideas have the potential for further development and which hypotheses proved less productive than we initially expected.

It is also important that the festival operates with a limited budget, meaning that we had to find creative solutions for a vast space that was not originally prepared for exhibition purposes. I believe we succeeded.

Technology can quickly become the main attraction, while the artistic statement becomes merely an illustration of its capabilities.

What turned out stronger than your initial concept, what would you do differently today, and how did this experience change your perspective? Is there a plan for further development of Noah33?

This experience changed my understanding of my own profession. I realized that a curator needs to be more courageous and trust their own intuition. If there is an important theme, an inner necessity to explore it, and artists are willing to enter this experiment together with the curator, then it should be carried out.

What became stronger than the original concept was the direct contact with the audience. The viewers’ responses showed that the project succeeded not only as an experiment with AI-based curating but also as an independent artistic statement. In the future, I would like to make the internal mechanics of the experiment more visible, while avoiding turning it into a technological spectacle.
The research will definitely continue. In future projects, Noah33 will most likely continue to evolve as a model of relationships between human, machine, and collective consciousness.