Curation Is Creation: pegekama on FLUX, Agency and the Making of RIVEN
RIVEN explores reconstruction through fashion — garments assembled from fragments, raw seams, visible stitches and exposed structure. Black, ivory and grey textiles are held together by red thread, turning repair into ornament and imperfection into form. The series treats the body and the garment as one fractured surface: cut apart, reassembled, and made deliberately visible.
RIVEN is one of the twenty editorials in MALVIE MAIN ISSUE 72, the 190-page, perfect-bound November 2026 main issue of MALVIE Magazine.
Order MALVIE MAIN ISSUE 72 in print:
pegekama: the interview
Your recent work as a reviewer for the NeurIPS 2026 Creative AI Track brought questions of agency into focus. When you work with a generative system, where do you feel the real creative decisions happen, and what remains distinctly yours?
Reviewing work for the NeurIPS Creative AI Track made me think much more consciously about agency in my own practice. When you are evaluating how other artists and researchers work with generative systems, you inevitably start questioning your own relationship with these tools as well: Where does the system end and the artist begin? I don’t think there is a simple answer to that.
I don’t experience agency as something that sits entirely with either me or the generative system. It moves back and forth during the process. I usually begin with a fairly clear idea of the visual world I want to create — its atmosphere, palette, materiality and emotional tone. But generation introduces an element I cannot completely control, and I actually value that. The model can produce accidents, distortions or visual solutions I would never have designed consciously.
For me, the decisive creative work happens in how I respond to those possibilities. I generate, reject, redirect, refine and curate constantly. Sometimes an unexpected result changes the direction of an entire series; at other times I discard a technically beautiful image because it simply does not belong to the world I am building.
What remains distinctly mine is therefore not every pixel of an image, but the intention, the decisions and the context around it: what I ask for, what I pursue, what I reject, what I place next to what, and when I decide that a series is finished. The generative system proposes possibilities. I decide which of those possibilities become part of my work.
Your practice is moving from making single images toward directing complete editorial series, with casting, styling, composition and visual rhythm. How have your 25 years in graphic and editorial design shaped the way you build a coherent visual world rather than a set of separate images?
After more than 25 years in graphic and editorial design, I think in sequences almost automatically. I rarely see an image entirely on its own. I immediately see what could come before it, what should follow it, whether I need a close-up or a full-body image next, where the eye needs a moment of rest, and where I want to create tension.
This becomes particularly visible in the nine-tile grid that has become a recurring structure in my work. Before a series becomes a reel, the eight selected stills have to function together as one composition around the central title. That means casting, poses, framing, scale, direction and visual weight all matter. An image can be beautiful on its own and still not make the final selection because it does not contribute to the whole.
The same thinking carries into the reel. I don’t suddenly switch from still-image making to filmmaking; I animate photographs. The visual relationships between the stills already contain much of the rhythm, and movement, music and editing add another layer to it.
So perhaps the biggest influence of my design background is that I am always looking for a system rather than a collection. I want eight images to feel like eight different moments from the same world.
RIVEN was created entirely with FLUX. What did that model bring to the series that another tool would not have, and at what point did the choice of model become part of the creative concept itself?
I work with several generative systems, and I don’t believe there is one model that is best for everything. Each has its own visual tendencies, strengths and limitations. But for RIVEN, FLUX gave me something I had not achieved with other models: the ability to make an entirely imagined fashion concept feel physically real.
The series is based on deconstruction — raw black, ivory and grey fabrics, exposed seams, loose threads, pins and garments that appear to have been cut apart and reconstructed directly on the body. FLUX translated these materials with an extraordinary sense of physicality. The fabrics have weight and texture, the frayed edges feel tactile, and the garments interact convincingly with the bodies wearing them.
What impressed me just as much was the consistency I could achieve across an entire editorial. Models, poses, framing and silhouettes could change radically — from close beauty portraits to full-body compositions and multiple figures — while the underlying design language remained recognizable. For the way I work, that is crucial. I am much less interested in generating one spectacular image than in building a visual world that can sustain a complete series.
At some point I deliberately decided that RIVEN would remain entirely within FLUX. I wanted to see how far I could push one model and how much variation I could create without losing the identity of the collection. The choice of model therefore became part of the creative concept itself.
RIVEN ultimately changed the way I think about generative models. I increasingly see them not as interchangeable image generators, but as different visual instruments. And FLUX has become an instrument whose particular understanding of material, body and photographic realism fits remarkably well with the kind of fashion imagery I want to create.
Much of your process stays invisible: iteration, rejection, refinement and curation. What makes you set an image aside, and why is deciding what not to use so central to your work?
Creating the images is only one part of the process. What is often much harder is deciding which of them deserve to remain. I can spend a long time developing an image until it is exactly where I want it to be — and still reject it because it does not belong to the series.
I usually set an image aside when it repeats something I already have, when the pose or composition feels too predictable, when the visual language begins to drift, or simply when another image performs the same role more strongly. Sometimes the reason is difficult to articulate: an image can be beautifully executed and still disturb the rhythm of the whole. After more than 25 years of working with images, I have learned to trust that instinct.
Generative systems offer an enormous space of possibilities, which makes restraint particularly important. The fact that another variation is always possible does not make the decisions easier — quite the opposite. At some point you have to stop exploring possibilities and start making choices.
In the end, viewers only see the eight images that survived. They don’t see how many directions were explored, refined and eventually abandoned. But those rejected images are part of the work too, because every rejection makes the visual world more precise. For me, curation is not what happens after creation. It is part of creation.
You use generative tools to imagine garments, materials, bodies and editorial situations that would be hard or impossible to build physically. How do you keep that freedom from turning into imitation of conventional fashion photography?
I actually move between two approaches. Some of my work is quite close to conventional fashion photography, and I genuinely enjoy that. I love the language of fashion editorials — casting, styling, lighting, posing and the relationship between a model and the camera. I don’t think every generative image needs to justify itself by being physically impossible.
At the same time, I am very aware that AI can imitate work that photographers, stylists and entire production teams have spent years mastering. As a graphic designer, I am experiencing the same technological shift in my own profession, so I don’t take that lightly. It makes me particularly interested in exploring what generative imagery can add rather than only what it can replace.
That is why many of my series contain a twist that would be extremely difficult, expensive or simply impossible to realize as a conventional shoot. In NIVAL, for example, I placed swimwear models in an Arctic landscape surrounded by ice and glaciers. Other works turn garments into structures, organisms or materials that could not behave that way in reality.
So there is deliberately no strict rule in my work. Sometimes I want to create a fashion image that could conceivably exist in a magazine exactly as it is. At other times I want to use AI to push beyond the physical limits of photography. What matters to me is not avoiding the language of traditional fashion photography, but finding out how far I can extend it.
Perhaps that is the balance I am looking for: respecting where fashion photography comes from while exploring where a medium without the same physical limitations might take it.
With AICONS, you reimagine cultural figures as “Digital Memorials.” How does that idea of memory connect to your fashion work, and where do you see your practice going as these visual systems keep growing?
AICONS is the heart of my work and the one project I can imagine continuing for the rest of my creative life. In the past 18 months alone, I have created more than 180 AICONS, and the series keeps growing.
I call them “Digital Memorials”, and that idea has a very personal dimension for me. On my father’s side, I come from a long line of church painters. They worked within a tradition of preserving stories, figures and collective memory through images. I work with completely different tools and in a completely different cultural context, of course, but I find it deeply moving that images have somehow remained a way of remembering across generations in my family.
Emotional intimacy is central to AICONS. I am interested in iconic moments, but often even more in the moments that might have happened between them: quiet, private, tender or vulnerable situations that were never photographed. I am not trying to create historical evidence. These are fictional visual memories — an attempt to imagine the human being behind an image that may already be deeply embedded in our collective memory.
There is no master plan for whom I choose. I follow my instincts completely. Sometimes the trigger is an anniversary or something I have read; sometimes it is simply a face, a memory, a piece of music or an unexpected association. An idea can appear very suddenly, and once it takes hold of me, there is usually no stopping until the reel is finished.
AICONS has also begun to develop its own branches. With MYSTIC AICONS, I extended the idea from real cultural figures to characters rooted in mythology and collective imagination — Circe, Medusa, the Banshee or the Manticore, for example. In a way, the project moves between people we remember and figures humanity has been imagining for centuries.
But AICONS exists within a much broader practice. I use generative imagery for almost anything that makes me wonder what it could look like — fashion, history, portraiture, surrealism, cultural memory or subjects prompted by current events. For German Unity Day, for example, I created a reel about the period following the opening of the Berlin Wall in 1989; a few days later I might be constructing an entirely fictional fashion world.
I sometimes think of my profile as a magazine: you turn the page and suddenly enter a completely different subject and visual world. I don’t need every page to tell the same story. The coherence comes from the way I look at images and from how I construct, select and sequence them.
Perhaps that is also where AICONS, fashion and all the other parts of my work ultimately meet. Generative imagery allows me to move constantly between memory and invention, between things that existed and things that never could. As these systems continue to develop, I want to preserve exactly that freedom. AICONS will remain at the heart of it, but around it I want to stay completely open to whatever catches my attention next.
I have no idea whether tomorrow that will be fashion, history, music, a cultural icon or something completely absurd — and I rather like that.
RIVEN credits
Graphic Designer and AI Artist: pegekama @pegekama
RIVEN fashion credits
Images 1–10 — AI-generated work
Comments