Sébastien Cliche
- Anna Lilli Garai
- Jun 26
- 5 min read
Sébastien Cliche builds installations that reflect on the shifting relationship between digital tools and manual processes. In his recent work, he imagines a deserted office tower overtaken by a new kind of activity—part algorithmic, part physical. Using a mix of AI-generated content and hands-on construction, he creates layered environments where sculpture, image, and space interact. Office partitions, adhesive prints, and ambiguous objects are arranged directly in the exhibition space, inviting the viewer to navigate a system of visual echoes and uncertain functions. His approach is shaped by repetition, feedback, and a focus on the strange familiarity of forms that seem both designed and discovered.
Q: What first drew you to the idea of reimagining abandoned office spaces as speculative futures?
A: Most of my installations revolve around workspaces. In the past, I’ve staged surveillance rooms, laboratories, data centers, and film studios. I’m drawn to these environments because their aesthetic is conditioned by function, and they often show things in a kind of "in-progress" state. In a way, it’s like an extension of the studio.
The idea of working from offices and cubicles had been in the back of my mind for some time, and this new project—focused on tools—felt like the right moment to dive into that setting. It allowed me to bring manual and intellectual labor into tension.
Then, in terms of fiction, I had this vision of office towers abandoned by their workers, all of whom had left to work remotely. And I thought: what if that space, once full of people typing on keyboards, talking on phones or sitting in meetings, were now occupied by another entity? Some other kind of workforce, producing objects according to its own contingencies.
Q: You describe becoming “permeable” to the generative flow. What does that feel
like in practice?
A: I'm interested in unfinished or transitional forms, and in this sense, generative processes are at the heart of my practice. I explore all kinds of systems, whether performative, chemical, mechanical or digital. My use of generative AI is in line with this approach, placing me in a position of distance and indeterminacy. For this project, I developed a way of working that’s driven by feedback loops.
To explain it simply, I first trained a model using a bank of photographs of objects I had either created or selected myself. Some of the images generated by this model inspired me to create physical objects in the studio.
I then photographed these objects and reintroduced them into the AI in a more directed way to generate hybrid prototypes. These new images, in turn, led me to make more objects that I photographed, and so on. Throughout the process, I try to preserve my agency by staying intuitive and freely absorbing the shapes and materials the algorithm offers me.
What really interests me is bringing together two modes of production: one where the object is imagined through algorithmic experimentation, and one where it’s improvised in response to physical, material constraints. Doing both in parallel to build a coherent body of work is really exciting.
Q: How does working back and forth between AI and your hands-on studio process
shape your decisions?
A: I try to avoid working too systematically. What I mean is that I don’t break the process into fixed stages or assign AI to a specific phase. I’ve also reduced text-based prompting to a minimum, focusing instead on image combination prompts. This approach shifts the emphasis toward visual drifting rather than relying on language-based control.
One of the main challenges when working with generative AI is the exponential number of possibilities and variations that can flood the creative process. To avoid being overwhelmed by the thousands of options, you need to find a way to make quick choices to guide what comes next. In this project, that meant a mix of rational criteria based on tool topology, but above all, actions guided by intuition and a spirit of improvisation.
In the studio, I apply the same approach with physical objects and materials. Like a generative algorithm, I work with my own black box, deliberately engaging in a similar combinatorial mode — improvising new forms from a set of materials and principles of similarity.
Q: How do you decide when an object or image feels “real” enough to belong in your installations?
A: That's an excellent question. Everything is decided on a fine line. I need to feel that the image or object belongs to the world I'm setting up, so that it responds to an aesthetic territory that I've marked out myself. At the same time, I want to feel that the object could be external to me, as if it had somehow been found. I think this is when the object or image appears sufficiently real to me, taking its place in the open fiction that's gradually developing — when it feels plausible but slightly misaligned.
I should also say that having to match both generated images and physical objects turned out to be very practical. It helped me quickly discard AI images that felt too slick or over the top and just didn’t fit in.
Q: The blend of familiar and strange runs through the whole project. What’s your relationship to ambiguity?
A: Ambiguity is a deliberate strategy. It’s a space of agency for me and for the viewer as it resists the closure of a single interpretation. I’m interested in that tension: when an object evokes a use, a memory, or a system, but doesn’t quite resolve into any of them. It keeps the experience active, open, and unfinished.
In general, my projects aren’t based on external references like historical facts, mythological stories, or personal narratives.
Everything is there, in the installation itself. Of course, each visitor brings their own set of references that shape what feels familiar and help them imagine the function of the objects they’re seeing. But it’s not limited to an intellectual exercise; I’m very sensitive to form and materiality, as well as to the relationship between the body and space.
I also use recurring motifs to generate both conscious and unconscious networks within the installation, and to evoke a kind of déjà vu. By encountering similar shapes in slightly different contexts, the viewer can gradually piece together an overall impression and experience a form of immersion in the work.
Q: Do you think the idea of value is changing as creation shifts toward automation and resemblance?
A: The question of value had already shifted away from the act of making itself, moving instead toward curation, selection, and contextualization. The automation of certain stages of creation will likely accelerate this shift. At the same time, the abundance of cultural objects circulating in an attention flow driven by look-alike aesthetics seems to encourage both fragmentation and uniformity.
That being said, it is difficult for me to speculate too much, as things are evolving rapidly.
Still, I believe there is a fundamental value in the experience of the work that goes beyond the economy of attention and resemblance as seen on social media. It might sound romantic, but I am working toward a form of engagement that is certainly informed by context, but also involves a sensitive, active and reflective approach. This relationship with art requires time and investment, and in my view, it is through this kind of investment that value is truly created. And just to be clear, I am not talking about NFTs, ha ha!
