Ègor Kraft

 The following images of chairs were generated by a Stable Diffusion deep learning, text-to-image model. The prompt was ‘a single chair on a white background’. The resulting dataset was used to train the Stable Diffusion model again, expanding its knowledge of what an image of a chair on a white background is.
 This process was repeated until the 6th iteration, where a previously figurative image of a chair was interpreted as an abstract digital colour noise as a result of so-called data cannibalism. Thus demonstrating how susceptible Ai-based computational ontology to corruption through a process of feedback loop of data in an echo chamber of a model.