Comparative analysis of the uses of AI in comics between the two spheres of production

Based on classic works on the seriality of cultural works (Letourneux, 2017), I would like to examine what generative AI brings to comics. Based on various elements that characterise serial production (media, genres, collections, characters, transfictionality, transmedia logic) and the constituent properties of generative AI (annotated training corpus, generic or specific generation models, links between prompt and unpredictable result), I will examine the potential use of the tool for the creation of mass-market comics. I will attempt to answer the question: what problem does AI solve for comic book production? What service does it provide and at what cost (Alombert, 2025; Bender & Hanna, 2025; Carbonell, 2025)? What skills must authors and readers develop to use it? As a counterpoint, I will examine the practices of generative AI within comics in the sphere of restricted production (Dubois, 1978; Bourdieu, 1992). Using works produced by Manouach (The Neural Network, 2020-2022; Fastwalkers, 2021), I will show the gap between the discourse surrounding the desired and anticipated uses of AI and the effects produced by the artist's practice. The aim of this presentation is to reflect on the tension between the emergence of a new tool, its promises linked to the context of enunciation, and its current practices.