Linda Dounia on Memory Machines

Artist Linda Dounia spoke with Peter Bauman (Monk Antony) about the potential of AI to combat cultural as well as biological loss, using her project Flore Perdue as a springboard.
About the Author
Linda Dounia, Flore Perdue (Test Output), 2024. Courtesy of the artist and Bright Moments

Linda Dounia on Memory Machines

Artist Linda Dounia spoke with Peter Bauman (Monk Antony) about the potential of AI to combat cultural as well as biological loss, using her project Flore Perdue as a springboard.

Peter Bauman: With Flore Perdue, I see a project celebrating apparent opposites. It touches on conservation and technology, memory and imagination, grief and hope, organic forms and geometric compositions. How do these juxtapositions contribute to the overarching narrative of the project?

Linda Dounia: I think the idea of duality is strongly present in this body of work, from when it was conceived to this new chapter. It's not so much conceptually how I approach the work. It's more about how I feel about the subject. I'm using a machine; I'm using something that is not human to remember, which is a very human trait or capability. From the get go, I've had this idea of, “Is it a collaboration? Or is it a conversation?” There's a stark duality in trying to do something very human and very effective using something that doesn't necessarily have those capabilities. With AI today, even if we want to personify it and give it sentience, the truth is that it’s just really good at reading data and learning from the patterns within the data. Is it feeling? Is it capable of remembering? Is it capable of having nostalgia?

This project is, in many ways, about nostalgia, the idea of remembering things you've lost, and in this case, remembering things you've never seen. It’s about being curious about the past.

Peter Bauman: You asked whether it was a collaboration or a conversation. What do you think?

Linda Dounia: It depends. GANs feel very much like an extension of myself. So in that sense, it's almost like an amplifier. Or I like to use this word, griot, which refers locally to people who are responsible for passing on oral history traditions. They will, for example, ask for your last name and they'll trace your lineage all the way back. They have this capability of remembering and recalling information when you need them to that you're not necessarily capable of doing yourself. I like to use that word griot when I think about GANs because they do feel like a storyteller or an amplifier. It feels like it's telling my story in a way that I haven't been able to digest yet because I don't take that approach.

When I think about my experiences or my intuitions, like everything that I've ever made or thought about, I can't really look at it within myself. I can reflect on specific things, but I'm not able to map my brain and then spew out what I've learned. But that's, in essence, what a GAN does. It's capable of digesting all of that information, which comes from me, and then giving me new insights.

So if it’s like a griot, I don't know if that's a collaborator. I think there's still a very practical relationship happening here that, when you're collaborating, doesn't necessarily hold true. I don't feel like I'm using my collaborators, you know? When it comes to generative AI, it feels more like, “Hey, this is a query. Answer my query and do what I say.” So that doesn't feel like a collaboration or a conversation either. I think a conversation can occur where there's a true back and forth. This often happens when you have to combine mediums because the complexity increases and you're creating an environment where you have less control. The translation from one medium to another is completely unexpected or generates completely unexpected results. Therefore, you constantly have to ask a question again, fine-tune and reevaluate.

Linda Dounia, Flore Perdue (Test Output), 2024. Courtesy of the artist and Bright Moments

Peter Bauman: Lauren Lee McCarthy said something similar to me recently about AI being more like a servant than a collaborator. You mentioned that one of the project's themes was nostalgia and curiosity about the past. You've spoken elsewhere about how childhood memories strongly impact your work. Could Flore Perdue be seen as an attempt to recreate personal memories of the past as well as the preservation of nature? What traces from your childhood come through in the project?

Linda Dounia: When I decided and accepted that I was an artist and that's how I was going to spend the rest of my life, I realized that there was a draw to my past and my childhood specifically. The reason—I've thought about this deeply—is that there’s absolutely no way to get back to that time. Not in the sense that I can't go back in time and find it but in the sense that I can’t go to my old neighborhood and get the same feeling my old neighborhood gave me, because literally, architecturally, that place doesn't exist anymore. It's the complete vanishing of a physical space. The people who inhabited that physical space and the interactions they had with their environment and with each other—that's all gone. If you go to my hometown right now, where I grew up, Dakar, Senegal, it's a completely different demographic, completely different architecture. In the last 20 years, it's just transformed into something new. It's as if nothing ever existed there before. It's a total and complete loss. I can't even go on the Internet and find photos of where I grew up because there's not even much photographic evidence. It's not something people thought to document or put on the Internet.

I’m drawn to AI because it's the only way I get to—aside from in my own mind—materially recreate places from my childhood, channel those feelings and reflect on the interactions. That’s what made me. 

But I just can't show it to anyone. I can't even understand it myself to a certain degree until I see it and get transported back into that place. That's the philosophy of Flore Perdue because there was such a big presence of flowers in my life growing up. My recent fxhash project, Le Bouquet, actually commemorates a flower offering that my grandma made every morning to the people who'd left the family. It was an offering to the dead and she would usually use jasmine, but in the absence of jasmine, because it doesn't grow all year, she would pick different flowers. 

This project and all the other projects started with the question: What's left? What's left to show? There’s a way to see what we lost from just this simple moment—the mundane activity of my grandma. It has led to so much new work and many interesting questions.

Peter Bauman: That explanation highlights the significance of the flowers in the piece. They serve as the conduit between the real and the imagined, which is expressed through your technique of obfuscating certain details of the flowers—of these organic forms—with abstract geometric shapes. What relationship between the abstract and the organic are you conveying there?

Linda Dounia: In the first installment of the project, Once Upon a Garden, I didn't need to do anything to the images themselves, aside from using collage, because they were exactly as I expected them to be. They were spectral, ill-formed and weird-looking. You knew you were looking at something resembling vegetation but not really a flower. Your brain almost wants to believe it's a flower because there’s enough suggestion but it really isn't right.

But then AI evolved so much in the last couple of years. It's remarkable. From Once Upon a Garden in 2022 to today, there's been such a huge leap in photorealistic AI. So I wanted to take the old version, give it to this supercharged, photorealistic AI and see what happens. The results were staggering. You almost question whether it’s real. There's texture and organic imperfections. You can sometimes even see a little dirt on the plant or you can see that it's dried up in some spots. There's just so much diversity and beauty in how realistic it is. But I felt that it was wrong in the sense that memory is, at best, patchy. 

Your brain fills in the gaps. But there are gaps and that's the most interesting part about memory—you cannot recreate it and serve it as a full, finished thing. That's not how it works.

So at this point in the project, I realized I had to use code to try to remove the detail. I could have done a gaussian blur and just blurred out the flowers. But I wanted to contrast the photorealism with these massive dots to show that memory is spotty. It looks good; it looks real but it is not. I'm hiding some of the details, forcing the viewer to miss them.

Linda Dounia, Flore Perdue (Test Output), 2024. Courtesy of the artist and Bright Moments

Peter Bauman: You mentioned using AI to aid with personal and collective memory. Why else do you choose GANs and code to express yourself artistically?

Linda Dounia: It's a good question. GANs are a softer way to experience AI that feels a little less dictated by big tech. It's a different entry point to AI. If I were to just interact with AI by prompting, it would feel like reinforcing the paradigm that everything is a product and everything is a service.

I'm ultimately a customer all the time and everything is productized; everything is served to you. Ultimately, you're seen as an economic agent with purchase power or no purchase power and that's it. It's such a brutal way to be human.

You're in a perpetual state of wanting, not having, not being able to afford and comparing yourself. It just diminishes our humanity. I think prompt-based software fits in the same paradigm. It's for-profit companies that are making it and therefore they're building it in the same way they built everything else. For me, engaging with GANs feels different. I think they're an alternative narrative of how AI could happen, of how AI could evolve, where instead of moving in the direction of general models, you're moving in the direction of smaller, more specialized models. It’s like a mushroom type of evolution, as opposed to this very hierarchical way of moving and distributing power. GANs feel more suited to what my personality is when I look at tech. 

Code just fits my personality too. The more I learn about code, the more I learn about the industry and its evolution. There's a bunch of mavericks—a bunch of people doing random stuff and talking to each other. It's so punk. That just sits well with my personality.

Peter Bauman: Speaking of mavericks, in the Curator’s Note for the recent Feral File exhibition you curated, you wrote about being “defiantly visible,” being boldly and confidently present despite challenges or obstacles. Do you see that concept pertaining to Flore Perdue?

Linda Dounia: This is something I've always felt, right? I tweeted recently about John Friedmann’s Core-Periphery Model. Essentially, you can divide any locality by how far people are from accessing everything from basic necessities to luxury items. I always felt, growing up, that I was definitely on the periphery of something. My world didn’t look like this world I was watching on television and that seemed like a better world.

I think what's happening with tech is that it’s mostly developed in, quote-unquote, the global north. I think that’s doing a disservice to the tech itself because it's not considering the needs and perspectives of people outside of the core world from which it's conceived.

We also see it in tech-related industries such as medical research, which is not serving the female population well because it’s largely been tested on men. Technological waste is another problem, with literal waste being shipped to the periphery. It's making life more dangerous and carcinogenic, even though they don't actually have access to the technology when it's made. These are the ramifications of not considering this huge group of people. Now AI is unearthing the fact that even in terms of data, the people in the periphery are underrepresented, if not misrepresented. That's where it gets dangerous. 

You have to be defiantly visible in the sense that you must not succumb to the temptation of saying, “This technology is not made for me; I'm not going to use it.” You have to say, “I'm going to force my way into telling my story, saying what I want to say using this technology, even though it was not built with me in mind, and even though it has no context on who I am, specifically in the case of AI.” 

Because if I don't, I'm doomed to disappear. It's just so simple; it's scary. That's why I relate my childhood so much to what's going on with AI. There's a total disappearance and a total loss of my childhood, not just in my head but physically. I see the same thing happening culturally to my identity, not just as a black woman, that's a very big term. As a Serer, mixed with Jola, mixed with Lebanese, growing up in Senegal, there's a very specific mishmash of identities that I grew up with that, when I try to inquire or investigate, using AI is totally absent.

Any kind of visibility is defiant because it doesn't exist. And it's sort of countering this total obliteration that you'd be faced with otherwise.

Linda Dounia, Flore Perdue (Test Output), 2024. Courtesy of the artist and Bright Moments

Peter Bauman: These themes, like your childhood, relate to your artistic practice. But you were a designer for over a decade before saying to Artxcode that you “transitioned to becoming an artist.” I’m interested in this transition and what the difference is between being a graphic designer and an artist.

Linda Dounia: Personally, from a craft standpoint, I don't see a huge difference. From a market standpoint, there's definitely a huge difference. A main difference is the presence of a client, which you don't always have in the design process anyway. In design, I start the process by just being creative and having fun, which is how I start my art projects as well.

There's a huge difference on the market side of things. The markets decide what is art and who is an artist. When I first started, it was very difficult to see myself as an artist because there’s an unspoken set of rules that, if you don't meet them, you can't really enter this world. It's really, really difficult to make it in the art world. 

When you think about the African continent and all of the different sorts of crafts that exist within each country, you have to make a decision on what is art and what is design. It gets really murky because the West sees art as something to be contemplated. It's in a white box. It's precious. It's hung somewhere. People stand at a distance and revere it.

On the continent, at least in my ethnicities, that's not how art is treated at all. It's an object that is part of daily life. You have the masquerades, which we call Kankurang, that happen every year during the rainy season. And there's this beautiful rafia-draped [palm fiber] man who is there supposedly to protect children from evil spirits. There's a whole ritual around it and the outfit he wears is so precious. But there isn’t a huge distance between the work of art and the people. Because it's part of the community, it's something that is used in daily life. It's a ritual; it's tradition. Dogon masks that are actually used in masquerade festivals as well can be found hanging in the British Museum but that doesn't make them less intended for use and interaction. 

Then the line gets even more blurred when you go back in history for me as an African because what was considered art actually met more of the criteria of design. Today, these objects are considered to be art, a place in a museum hung away from the communities that are meant to use them. That's sacrilegious in a way because a definition of art was imposed.

Peter Bauman: That’s a great point and I’ve always been fascinated by groups that have historically blurred these lines between the “white cube” art world and everyday life, going back to Gutai in Japan and GRAV in Paris. And I think NFTs enable us to carry on that tradition. They facilitate the reintroduction of art into our daily lives. Art is on our phones; it's the wallpaper on our computers and it's on our large TVs all the time.

Linda Dounia:
Yeah, our masks, too. We use them as PFPs too. People have been doing that for thousands of years. So we're almost going back to a broader definition of art, which is a long-time coming and important.

Peter Bauman: I love that point. Finally, perhaps my favorite individual component of your work, besides the complex emotional layers it expresses, is your striking and refreshing use of color. Where do you draw inspiration for your palettes?

Linda Dounia: If you were ever to google “Senegal” or even do street view and just zoom into the people, they usually wear garments that have a lot of different colors. That's my first palette inspiration. I will literally see my grandma or my mom wear something, take a photo of it and make a palette out of it.

Peter Bauman: I love how you’re now blurring these art-daily life distinctions, bringing the palettes of everyday life into the art world.

Linda Dounia:
It’s from everyday life that I always find these fantastic combinations because they’re colors I would never think could go together. And you cannot use color theory. You cannot, because they defy it. Sometimes you'll have the greenest neon on a very close color, like yellow. The combinations are so refreshing. If I were to go on the Internet and look for color palettes, it would look gray and dull in comparison. Oh, my God! Designers, we hate color. Apparently, it's just getting grayer and grayer. So I have to look at everyday life and the organic to get my colors. It's outfits; it's plants. It's from the pictures I take of people or buildings. All of these inform my palette. So thank nature.


Linda Dounia is an artist and designer investigating the philosophical and environmental implications of technocapitalism.

Peter Bauman (Monk Antony) is Le Random's Editor-in-Chief.