Fuse by Scott Draves

This moment stands as one of the earliest predecessors of the big-data, machine-learning AI art of today, as opposed to the symbolic AI art of Harold Cohen. “One of the earliest machine learning image2image algorithm[s],” according to Draves himself, Fuse stands as a critical step in the development of AI art. Scott Draves developed the Fuse algorithm, drawing inspiration from the Dissociated Press and Cut-Up technique. The process worked by slicing and rearranging images into new compositions, predating Efros and Leung’s texture synthesis (1999) and Robert Silvers’s Photomosaics (1995). Released as open source in 1993, Fuse became influential in digital art. Draves' self-portrait, composed of fragmented images, exemplified its creative potential and is one of the earliest examples machine-learning art as we know it. Draves’s work was inspired by the artificial life work of Karl Sims and also can be traced back to the work of William Latham.