Generated tools: A Defamiliarizing Approach to Creating ML Art

SHORT PAPER. In recent years there has been an increase in the adoption of machine learning (ML) systems that can generate novel images. This increased use may reveal the beginning of a familiarity in which the implications of these emerging technologies are naturalised or made increasingly invisible. Thus, practices which can disrupt familiarity may allow us to create experiences of heightened awareness in which we can consider our engagement with this emerging technology. In this paper, I discuss the outcomes of working with a Generative Adversarial Network (GAN), using a dataset created from the hand tools section of a popular Australian hardware store. Through this creative practice, I investigate how artists can use ML as mechanisms for creating artworks that disrupt, investigate and defamiliarize the known.