From oral tradition to generative AI

regimes of authorship, archive, and responsibility

Abstract

This article examines the contemporary reconfiguration of authorship in the face of generative AI through a material and institutional genealogy of written culture, articulating regimes of discourse production, archive, and responsibility. In recent scholarship, AI appears neither as autonomous creativity nor as a replacement for the human author, but as a device that redistributes the authorial function across technical infrastructures, corpora, invisible human labor, and academic and legal validation mechanisms. Drawing on Milman Parry (1971), the article shows that in primary oral cultures narrative operates as collective traditional performance, without fixed text and without modern individual authorship. Writing, by stabilizing the utterance, creates the conditions for separating work, narrator, and organizing instance. With print culture and the editorial economy, authorship also becomes a technology of control: classification, authentication, accountability, and property. These historical transformations help explain the modern constitution of authorship and its relations to textual fixation, editorial circulation, and legal regimes of discourse. In parallel, generative AI produces effects of authorship through pragmatic simulation and probabilistic recombination of the archive, challenging originality and displacing responsibility onto human chains and institutions. Finally, the article discusses NotebookLM as a form of “expanded reading” that intensifies operations of curation, indexing, and synthesis, reinforcing authorship as process governance and an ethics of validation.