From Representation to Prediction: Theorising the AI Image
摘要
This chapter explores the emergence and implications of AI-generated media, with a focus on images. AI generative media is situated within historical context, drawing connections to earlier artistic movements and techniques like photomontage, postmodern bricolage, and database art. The analysis emphasises how AI continues the tradition of creating new art from accumulations of existing cultural artifacts but does so through novel methods of pattern extraction and prediction. The chapter examines several key aspects of AI generative media, including how it enables new forms of “media translation” between different types like text and images. I analyse the tension between stereotypical/idealised outputs and unique human creations, noting AI’s tendency to generate less unique and specific details in many cases. The relationship between subject and style in AI-generated images is also explored. I demonstrate how engagement with AI generative tools offers new perspectives on longstanding questions in art and media theory. AI generative media continues the tradition of artists expanding creative possibilities through emerging technologies while also challenging our understanding of style, content, and the nature of artistic creation.