Beyond creativity: Decoding changes mediated by Generative AI models in visual generative media and creative production

Abstract

The paper explores the use of generative artificial intelligence (GenAI) in creative production of images and videos. We focus on Runway, a US based AI company developing multimodal foundation models geared toward video editors and VFX artists. We borrow from science and technology studies, critical data studies and media studies to examine how GenAI technologies are mediating creative production, idealizing (affordances of) tools, and disrupting creative workflow. We pay close attention to embedded assumptions about the users, design choices, and technical constraints in the creative production. Using the app walkthrough method, we analyze the temporal and technical aspects of the creative workflow as mediated by Runway’s Gen-3 Alpha model. Our research shows three dominant ways in which Runway conceptualizes creative production: immediacy (blurring ideation and execution); iterativity (prompt engineering); and control (agency as transferred from creative thinking and doing to aesthetic judgement). In this way, the model emphasizes creativity where linguistic interaction, via prompting, relies on continuous exercise of aesthetic judgment constrained by the technical system (source datasets, algorithms, parameters); it shifts the locus of creative production by widening the gap between thinking and doing in creative practice. Hence, these “new” technologies propose moments of rupture wherein logics of industrial productivity, extraction, and automation are reinforced, and creative expertise, skills, autonomy, and professional identity of creative workers are reshaped.