The Death of the Author in the Age of Algorithms: Intertextuality and Surrealist Hallucination in AI-Collaborative Generation Z

Abstract

Generative AI now sits at the writer's desk as a co-author, and the arrangement unsettles long-standing assumptions about textual ownership, stylistic originality, and the source of literary voice. This study reads four AI-collaborative novels by Generation Z authors through Julia Kristeva's notion of the text as a mosaic of quotations and Mikhail Bakhtin's dialogism, combined with qualitative content analysis. The corpus consists of novels co-written with GPT-4, Claude, Gemini, and LLaMA. The analysis draws on close reading, documentation of prompt histories, and systematic coding of linguistic anomalies across the four texts. Six recurring intertextual patterns emerged from this reading, along with a new taxonomy of five hallucination phenomena that work as surrealist stylistic devices in their own right. Temporal dissonance, ontological ambiguity, and semantic slippage produce a form of machinic intertextuality that complicates Roland Barthes' concept of the death of the author: authorial agency survives, but split between the human prompt and the algorithmic response. The study offers a framework, grounded in close textual evidence, for reading AI-collaborative literature, and argues that hallucination, read on its own terms, functions as the signature rhetorical mode of algorithmic co-authorship.