AMANDA WASIELEWSKI

Image generated by Amanda Wasielewski.

Afterimage by Amanda Wasielewski:

I’ve generated a lot of AI images for both my research and also my art practice, it is driven by curiosity and exploration. I created this image about a year and a half ago, and I keep coming back to it. The prompt was inspired by the famous line from The Songs of Maldoror by Comte de Lautréamont: "…as the chance juxtaposition of a sewing machine and an umbrella on a dissecting table!" This line was crucial to the Surrealists because it spoke to the unexpected chance encounters of objects, and I wanted to see what an AI tool would generate from that.

What intrigued me was the unexpected appearance of a weird object on the left side of the image. The sewing machine and umbrella are represented here, though not quite as I envisioned, but then there’s this other object. It is the most uncanny part of the image. It looks like some kind of medical equipment stool covered in periwinkle, shiny fabric, something you’d expect a surgeon to wear in an operating room, but it has a metal pan on top. I kept coming back to this odd object, wondering what it means, why it’s there, and how it was conjured. It feels like fragments of associations, visualized through the AI model based on the prompt. For instance, I wanted an operating table or surgical table, but what I got was something closer to a sewing table. Still, the surgical part comes through with this strange figure that seems to have wandered in from the side.

I was thinking about what kinds of pixel fragments or textures the model associates with specific words. It’s a mistake, but it’s also evocative in this weird, uncanny way. The more you look at it, the less you understand. I could point out other absurdities in the image, like the umbrella being both a rain umbrella and an outdoor patio umbrella, or the lamp hanging from nothing. These are weird, but not inexplicable like this stool which feels more unknowable.

I’ve generated many different images, and there’s always something interesting that comes through. I think many commercial AI tools are refined repeatedly to become more and more standardized, so users get what they expect and no longer encounter strange, unexpected results. The commercial tools don’t want weird stuff. But in terms of an art practice, what’s most interesting to me are the weird things—the fragments, the mistakes, the artefacts of what's happening within the model to generate these images based on whatever associations the words in the prompt carry.

There’s something about how we want to trust and believe in images, no matter the medium. We’ve had years of easy image alteration—from Photoshop to social media filters—and yet, we still want to believe in what we see. I look at these images, and it’s so obvious, how could anyone possibly believe in them. But I think we still want images to communicate truthfully, and that’s something I find interesting. Over time, those images will become less creepy, and I think that will cover up or beautify the weirdness. But right now, we’re in this moment in image culture where we can still see the oddities before they’re smoothed over.

Wasielewski is Associate Senior Lecturer of Digital Humanities and Associate Professor (Docent) of Art History in the Department of ALM (Archives, Libraries, Museums) at Uppsala University. Her research focuses on the use of artificial intelligence tools to study and create art and images.

Afterimage is an ekphrastic series about that one image you see when you close your eyes, the one still lingering in your mind. We invite artists and writers to reflect on an image they can't shake. This column has been a part of Objektiv since our very first issue in 2010.