Each Untitled Film Still image is a frame from an imagined film. The generative adversarial network (GAN) machine learning models used for Untitled Film Stills are capable of producing high-fidelity, photographic images, but that’s not the direction of this work. Each Untitled Film Still is an uncanny picture from a dream; it is evocative of the history of oneiric cinema. The images oscillate between representation and abstraction; they require interpretation to make meaning.