Abstract

This study identifies alternative models for the production of AI-generated images to those currently used by mainstream AI platforms. Based on primitive computational art processes, these systems allow designers to gain greater control over the final visual result while avoiding potential issues with intellectual property theft and breach of privacy. The article starts by analysing the level of artificiality that might be effectively attributed to each part of the creative process involved in the development of AI-generated images. It then moves on to discuss the extent to which individual actors intervene in each of these stages, with the aim of untangling the different subjectivities that would ultimately shape the meaning of AI-generated photographs, and identifying existing AI models that are able to offer practical solutions for every designer to retain creative autonomy over their AI creations.