In the summer of 1895, Freud had a dream about a former patient, now known as “Irma’s Injection,” that helped him understand the role of the unconscious in dreaming. According to Freud, our unconscious constructs dreams to express our deepest wishes and then distorts those expressions to protect us. By this logic, a dream is therefore a form of wish fulfillment and suppression.
Although Freud’s theories have long since been rejected, they provide a useful framework for thinking about why dreams and the unconscious are so frequently invoked to describe artificial neural networks and their psychedelic visualizations. In the same way that one never fully understands a dream’s latent content, artificial neural network architects do not fully understand the mechanisms of deep learning. Does this lack of understanding act as a form of dream distortion, shielding us from our deepest wishes? When we build technologies that “dream,” what wishes are we trying to fulfill, and why must we protect ourselves from fully expressing them?
Artificial neural networks are modeled after our own and trained on the internet of images we have produced. Given that these models “see” with the same biases of mainstream culture, what are so-called “neural net dreams” telling us about our own dreams and unconscious desires?
Created using a deep learning text-to-image model, Fulfillment of a Wish experiments with storytelling in latent space, reflecting our own vibrant yet nightmarish images back at us.
