Abstract
AI image generators gained widespread public attention in 2022, representing a leap in creating novel visual content through pattern synthesis from massive datasets. As these tools became more widely used, several lawsuits have been filed by creators against AI image-generation companies—including Stability AI, MidJourney, and DeviantArt—alleging copyright infringement. Central to these allegations is the assertion that these AI models can ‘mimic the plaintiffs’ artistic styles’ without authorization, presenting a breach of the venerable ramparts guarding traditional definitions of copying and representation (Andersen v. Stability AI Ltd. 2024,4). These claims center on “style transfer,” an emerging deep learning technique introduced in 2015 that enables neural networks to separate image content from style elements and recombine them, applying one image’s style to another’s content. While media scholars largely viewed the technique as operating within reproduction, aestheticization, and automation, building on the work of scholar Mario Carpo and literary critic N. Katherine Hayles, this paper examines how these copyright claims reveal a deeper philosophical problem: not only the examination of how generative AI resurrects antiquated theories of artistic creation that the creative community deliberately excised from modern discourse on art and design, with abundant justification, but also as an invitation to explore how after two millennia of confining mimesis solely to human domains, we now need to confront its recasting as it is practiced by intelligent systems beyond human cognition—a reconceptualization that will transform existing concepts of originality, authorship, and transformative use through which these copyright disputes are adjudicated.
Presenters
Maayan AmirResearcher, Director and Curator of University Galleries , Department of the Arts, Ben-Gurion University of the Negev, Israel
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
Creative and Cultural Technologies
KEYWORDS
Mimesis, Style Transfer, Non Cognitive, Generative AI
