Human x AI
Liliana x Dall-E
During my high school days, I began to have a dream where I would chew gum, but when I tried to spit it out, it would turn into a never-ending strip that I couldn't completely remove from my mouth. The dream would continue, and I would struggle to extract the gum until I woke up. I could never entirely remove it, and some bits would always get stuck to my teeth. Soon it became a recurring dream, sometimes repeating itself several times a week and other times several months apart.
Later in my life, during therapy sessions, this dream became central to defining my traumatic childhood. It was interpreted in many different ways: an inability to express myself, repressed emotions, attempts to eliminate or fix relationships, anxiety, frustration, and even a far-fetched idea that my mind might have been signalling that I had scar tissue - because of the visual resemblance between scar tissue and chewing gum - within my body that would eventually cause life-changing surgery years later.
Regardless of the reason for the dream, what persists is the feeling I have: a sort of muted nightmare, where there is no fear or anxiety but rather an eternal punishment which I diligently try to finish, stuck into a Sisyphus scenario. What draws me into thinking about it, is the quality of our unconscious to gather past experiences, memories, and emotions, and to use this information to create new symbols. Similarly, I have been intrigued by the new AI technology, and in particular imagery creation algorithms, that can draw upon vast amounts of data to create new imagery. The analogy between the process of AI and human dreaming has pulled me into using the new technology and creating a visual representation of my dream.
The process was a complete departure from my usual photographic approach. While I might plan what I want to capture in a photograph, there is also an intuitive process where I have to see, observe, look, capture, and frame. Possibly an unconscious choice between one moment and another is also part of it. However, working with AI technology, I had to start with a text description such as "a kneeling woman pulling infinite chewing gum out of her mouth," which serves as the visual concept or idea to be translated into an image. DALL-E then utilises its deep neural network architecture, trained on a massive dataset of diverse images, to generate a corresponding visual representation. Instead of constructing an image or capturing a moment, I had to find a visual language that the AI would interpret into what I wanted to see. The AI model goes through a series of iterations and refinements to produce an image that closely aligns with the provided textual description. It analyses patterns, shapes, and contextual cues from the text, and combines them with its understanding of various visual elements acquired from the dataset.
What grabbed my attention was an image where the imperfections of AI imagery were visible: an amalgamation of skewed elements that overall resembled my description but was not a perfect copy of reality. The type of imagery that you know it’s created digitally, but which was closer to translating the feeling of the dream: stuck and embedded, struggling, confused, disturbed, scrambled between symbols that the AI put together. I would say, not very different from what my brain does during the night, feeding me back mixed imagery from all the experiences I have had. The chewing gum became part of the body and the body became an amalgam of chewy, sticky pieces: the never-ending struggle melting into a disturbing coexistence of elements. The body becoming the chewing gum itself.
The imagery already comes with a title: Human x AI (Liliana x Dall-E). A sort of collaboration between myself and the AI; a co-creation. Nonetheless, the collaboration extends to millions of other people that have created the imagery that was fed into the AI, and from which the imagery has been pieced together. A collective portrayal and interpretation of my dream.