Dream Field (2024) is an extended reality experience that explores new opportunities for world-building afforded by digital rendering technologies: game engines, machine learning, and generative imaging.
In Fang’s fictional universe, an AI with self-determination has created its own world—a simulated version of the earth where the AI can finely control the virtual ecosystem to maintain ecological balance. Through a real-time video feed, the audience’s faces are translated into virtual creatures. By stepping in front of the AI’s lens, humans can see what their more efficient interspecies role would be in this idealized environment. Blurring the semantic divide between man, machine, and environment, Dream Field invites viewers to confront their own roles within ecological systems.
Throughout her practice, Fang experiments with non-conventional mediums for worldbuilding. In Dream Field, real-time AI image generation continuously evolves, updating the displayed scene constantly. The scene itself is built on top of an open world made in a traditional game engine (Unreal Engine 5, TouchDesigner). A USB webcam captures audience members, porting them into Fang’s AI-driven world. By embedding prompts into the invisible infrastructure running the blended 2D/3D visualization, Fang creates a hybrid space for play, experimentation, and reflection.
Overall Dream Field has an ethereal, disorienting effect. Warped aesthetics obfuscate processed imagery from the camera, rendering the scene difficult to read. The abstract qualities of Dream Field make reference to the concept of “hallucinations,” or the “ghost” in the machine: in popular culture, this is how unexpected outputs from machine learning systems are described. The resulting world that we see through human eyes is an idealized vision—a dream of reality—that echoes the emotional experience of ecological grief.
