Benji Friedman is an artist and software developer who works across generative AI, photography, and digital media, treating each as a system to be built and then explored for what it produces.,Based in Berkeley, California, he spent most of the 2010s making abstract paintings and collage before shifting his practice toward AI image synthesis, using pixel manipulation, randomness, and controlled chaos to create images between intention and accident.
Friedman describes his AI-based work as synthography, a term coined by Elke Reinhuber for image-making built from the orchestration of generative models rather than a camera or a brush.Much of that work turns on CFG (Classifier-Free Guidance) scale, the parameter that controls how closely a diffusion model follows a text prompt.At very low settings the model drifts away from the prompt and generates from its own internal patterns instead, producing what Friedman calls proto-images.,He has produced two such bodies of work in this mode, an original Low CFG set in 2023 and a further Low CFG 2025 series, treating the curation of what to keep from hundreds of generations as the real creative labor.
Friedman's practice centers on working in series rather than single images, since repetition across dozens or hundreds of variations is what reveals a generative concept's edges and exceptions.His Plant People series explores the boundary between human and botanical forms across sixty-five variations,while his Lamassu series reworks the ancient Assyrian guardian-figure archetype,and his Forest Mazes series merges forest imagery with labyrinthine structure.
