29 Jul Using my toolkit of power to create dialogue with the public
Sahar Sajadieh is a computational performance artivist and researcher who works at the intersection of creative, performative, and ethical AI, digital performance, and socially engaged design. Sahar has recently completed her Chancellor’s Postdoctoral Fellowship at the University of California, San Diego, Department of Visual Arts, and is currently a fellow at the MIT Open Documentary Lab.
An interview with Sahar Sajadieh by Karen Cirillo
What was your research question?
I originally came in with a question about disinformation that arose from an artwork-in-progress I created with a group of female artists, activists, and scholars, interrogating information clusters and social media echo chambers. We created characters and trained them on different social media platforms with different agendas. We examined how these groups, representing opposite sides of the political spectrum, were recruiting followers and how their specific visual identity, as well as the structure and color palette used in their posts, served this purpose. We then created a playful, performative artwork based on these findings.
When I started my fellowship, it was around the time the Women, Life, Freedom movement was happening, which was deeply core-shaking for me. I spoke with many other Iranian artists at that time, especially female artists in the diaspora, and we all felt so much pain in our hearts and our throats, feeling that we needed to do something, but we couldn’t breathe, we couldn’t speak.
At the same time, as artists, we felt a great responsibility toward our art, the inspirations we were feeling and the urge to create something. The original question of disinformation shifted toward this issue of human rights, toward the pressure we were feeling, and how to use the creative, computational media that we have access to as a tool for creative response.
Many Iranian artists were performing and creating works at that time, but they were often covering their faces and identities. The question of visibility and invisibility was a central question. We wanted to be visible, but at the same time, we had to be invisible enough to feel comfortable practicing our art. And that’s when computational media, particularly generative AI as a co-creator, came to my aid in creating this safe space—a playful, performative, and poetic space. It also had a therapeutic effect, helping me express myself in a very different, unfamiliar yet liberating way.
I would write things—lines, statements, stories, poetic gestures—then use self-coded generative AI to craft images based on the texts, and then feed those images back into the system to foster the conversation and a new ecosystem of co-creation with generative AI. I would respond to those images by writing on top of them, modifying the code, and put them back into the system. A truly interesting creative loop emerged, opening a new world of possibilities for visual and poetic storytelling for me. And that became my vision for my fellowship at ODL.
What form did that take, and how did your background influence that?
What I ended up creating in ODL was more of an artwork and computational platform, which was my response—computationally, performatively, poetically and artistically—to the events around us. Becoming Schizophrenic is a broader umbrella for co-creation with generative AI, and my objective was to develop that platform and to create artworks using it.
I double-majored in computer science and theater during my undergraduate program, a time when this type of transdisciplinary practice in computing and performance was not yet common. Then I pursued my Master’s in performance studies, which led to performance theory and experimental performance becoming a more significant component of my research and practice. I completed my PhD in Media Arts and Technology, combining critical media and performance theory with arts and digital performance, computational media art and interactive installations, using coding as my craft and developing and utilizing various digital media platforms as my creative tools.
“Becoming schizophrenic” is a concept developed by Deleuze and Guattari. It’s not referring to schizophrenia in the medical context, but a term I find useful in describing co-creation with Gen AI. For me, it provides a model for this ecosystem of co-imagination and co-envisioning with a technology that is partly a reflection of people’s collective memories—good, bad, and ugly—and outputs images that resemble hallucinations. When you play with that, it can break many structures and narratives, and assemble things in new ways that are fascinating, but again, can sometimes be problematic. And that was the space I was exploring. I used the concept of “Becoming Schizophrenic” as a framework to help describe how this space—its aesthetic and the co-creation process—would feel and look.
How can I co-create with this technology, a technology that is highly problematic? I wanted to repurpose these computational systems—largely developed through funding from the military, state and tech industry, often for political and financial gain—and transform them into means for creativity, political and social activism, and therapeutic reflection in response to the events unfolding around us.
What kind of pieces did you create?
Tales of the Hair was my creative response to the Women, Life, Freedom movement in Iran. I developed my code, basically as a toolkit on top of Stable Diffusion, which is an open-source text-to-image generation platform. Using coding, these images, and my poetic creations, I made a video artwork that responded to the movement through art, performance, poetry and Persian mythology. My goal was to reclaim the narrative of the Iranian movement. I wanted the work to be something that would show the beauty of Iranian people, especially the women, in a time when the Western media had created a boogeyman out of Iran, and instead would truthfully portray the Iranian woman not as a victim but as a powerful and fearless force of nature. I wanted to showcase the power of Iranian culture, history, poetry, and literature through this work. These are the things that have shaped our culture and have made us who we are. And when the media censors and manipulates that full image, they’re simply feeding the public a reduced and distorted image that is intended to scare people. My work was not “political”; it was artistic and very personal and Political (with capital P) and, therefore, very universal. It was about women all over the world.
You talk about co-creating with generative AI, and you’re also talking about Western media’s stereotypes and misinterpretation of people and the diversity of people. One of the biggest problems with generative AI is that it’s reflecting off of a dominant language model that speaks mostly from a Western white context. So what is that process like, co-creating with generative AI that in its base nature is just replicating all of the things that you’re actually trying to fight against?
This platform has a lot of issues. As artists, we need to be very careful about how we use it, how we fine-tune and train it, and be aware of the sources we use (right?), who is creating it, who benefits from it, and what type of data it’s being trained on. When I first started using it, most of the platforms were trained on data from before the Women, Life, Freedom movement. So, the system had no idea what that even meant and wasn’t trained on more recent terms and events that I was referring to. I was getting information and images based on the memory of this machine, very biased and problematic, but not even current enough, as if it were this old grandparent who only remembers moments before certain dates and has all these discriminatory viewpoints.
For example, hair was a very iconic part of the movement. But I would get all these stereotypical images, even when I was prompting with “Iranian women cutting their hair” and “ripping apart their scarves”, images such as women with a full hijab and some random scissors on the side, which was absurd.
That’s why I used Stable Diffusion, an open-source platform that would give you much more control if you knew how to code. When [computational media and AI] were displaying problematic, discriminatory and stereotypical images, I needed to understand how and when they were generating them so I could find a way around it, create my own system, train it on my own data, and then start building this alternative ecosystem that existed in a different space. We need to constantly look back to review and see where we are standing, where these biases are coming from, and at the same time, equip ourselves with the skills needed for the application of these media to gain more control and power in this space.
I think it’s very important for media artists, performance artists, or even creative artists in any field to know some coding, because it gives them power. It gives them the power to look at the world, the media, and the way our data is being used and exploited, and then to find a way to view it from their own perspective, analyze this data, and creatively use and critique it.
How do you feel your time at the Open Doclab impacted this work?
It really was positively impactful. The first time I had to present, I wasn’t sure what I was doing and was worried about even talking about these issues [related to the Women, Life, Freedom movement], but our lab was truly amazing, first and foremost in terms of the people who have been part of it: a brilliant collection of artists, activists, filmmakers, academics, scholars, thinkers, and theorists. They were so supportive in helping me clarify my thoughts by asking me many new questions. I felt I had a community that was backing me up, one that was very sympathetic and supportive of the struggles I was going through at that time in a non-judgmental way. Having this ongoing conversation about my work and about other fellows’ was very helpful and therapeutic, but also thought-provoking. It was also helpful that I got the chance to meet many amazing scholars, artists, and technologists on the MIT campus. I also found one of my collaborators there, Manaswi Mishra, a PhD student in the Opera of the Future Research Group at MIT Media Lab. He’s an audio designer and AI engineer who worked with me on creating the sonic aspect of Tales of the Hair.
I wanted to ask you about affective storytelling. How do you define that?
It’s part of my larger question about human interaction with technology. It’s about creating works that lie in our intimate space; and sharing such a space with a machine. Can we develop an emotional connection with a machine to the extent that it becomes therapeutic for us? Can we possibly do that with systems built on binary structures—systems that are Politically problematic and inherently biased? [Other artificially intelligent systems] have existed before, but now there is a sudden explosion of AI that has opened up our human society to new forms of interaction with these algorithmic systems. I wanted to study the emotional relations in this emerging techno-social environment, both by creating technological performers that have affective interactions with us and by examining the effects that the layer of technology will have on this intimate, human-specific realm. Can we utilize these technological agents and emotionally charged spaces to tell stories and raise questions about societal issues?
Affective storytelling, for me, is a particular approach to telling stories—in both the process and the final product—and what it evokes within me and in that space between myself and the machine. I bring something emotional into this intimate space; images are generated, some resonate more deeply, and I feed that emotion back into the loop. That’s when I began developing a new watercolor aesthetic, which lends itself beautifully to visual poetry and storytelling. In every part of my artistic response—from the process itself, to the way the story unfolds, the poetic co-creation and co-imagination with the technology, and the final “product” (which is never truly final, but always a work-in-progress)—there are deeply personal, emotional, and communal elements embedded in both the subject matter and the form of expression. That’s what I refer to as affective storytelling.
It goes back to what you said about having these different backgrounds – computational arts, social design, performance… They all seem very separate, but they are all working together, and that fusion is what allows you to do this kind of storytelling. You’re bringing so many different influences into it that it actually becomes another thing entirely, which is both of you and also not of you.
Absolutely. I felt that using [generative AI] technology was something that was not of me, so it altered the shape of my expression, my medium, the responses and the type of artwork I ultimately created. I felt I was actually dancing, performing and writing poetry through this new medium, with a new visual language.
AI-influenced art can be a bit technical and digital and conceptual, but you are describing using technology as actually a way for you to be creative that you couldn’t be using traditional or more tactile methods, which is an interesting juxtaposition. How did using the technology make you feel freer and able to experiment more?
Art needs to touch something within me, but if it doesn’t hit that note, for me, it’s not an impactful or affective artwork. My work and interests lie at the intersection of all the disciplines I’m involved in, and one of my focuses is on embodiment and the various layers of identity that we encounter, these multiplicities of being that are constantly changing and in the process of becoming. My doctoral research focused on affective interactions with technology and the concept of liveness, specifically how liveness and presence change when the human performer is removed from the equation and replaced with technological bodies. What happens to presence and liveness when there is also a layer of technology in between me and the other person? So, I examined these new bodies, this new assemblage of human and machine, and the new forms of interactivity.
What’s next for you in the fellowship, and beyond?
I’m approaching the end of my fellowship. I’m still working on Tales of the Hair, planning to create different iterations of it, including an immersive installation with spatial audio and video projections on textiles. I’m also further developing AI-based computational tools for various types of storytelling, poetic creation, and art activism.
I’m also working on a project that started in my fellowship at ODL with another artist and director, Mathieu Pradat. We created an interactive augmented reality artwork, called the Paper Boat, as part of Layers of Place, a site-specific public AR Exhibition by the OpenDoc Lab, on the MIT campus until June 2025. Paper Boat draws attention to the rising sea levels, which are often invisible to many of us, and introduces a new layer to the MIT campus to visualize and provoke dialogue about climate change there. At specific locations, viewers can use their phones to scan manhole covers and trigger the AR application we developed. Virtual water will burst out of that manhole, covering the space and surrounding the participant. The objective is for them to rescue a tiny virtual human figure by taking him safely to a virtual paper boat while avoiding all the garbage and plastic waste on the virtual water. At the end of the experience, live climate data from NASA appears on the screen. The longer it takes for the participant to play, the more challenging it becomes, as the virtual water level around them gradually rises over time until they are fully submerged and lose the game.
Do you feel that art can be an activist tool?
Our job as artists is to look at the world and respond to what is impacting us, our communities, and our global human society; and to look at the past, look at the future, look at the broader spectrum of disciplines across different times and societies, and respond to them. My work is highly transdisciplinary, and I use a wide range of toolkits. The final format can take different forms, but the main thing that matters is what issues I’m looking at, what issues are inspiring me,
I use my toolkit of power—which is my art, my computing, my poetry, my performance—to not only respond to them, but also create dialogue with the public so they engage in a playful interaction that makes them think and potentially look at the world differently and create new types of interaction with other people and their environment.