The Death of the Reader 2.0: From Roland Barthes' Scriptor to the AI Prompter

Abstract

Roland Barthes' declaration of the "death of the Author" in 1967 appeared to dismantle one of the central hierarchies of modern textual culture: the belief that meaning originates in an autonomous consciousness standing behind the text. Yet Barthes did not remove the human subject from the communicative circuit. He transferred sovereignty from one end of it to the other. The Author disappeared, but the Reader was born as the privileged destination in which the plurality of writing could finally be assembled. This article argues that generative artificial intelligence produces a second and potentially more consequential displacement: the death of the Reader as the necessary endpoint of textual processing. Large language models do not merely generate linguistic material without a traditional human author; they retrieve, compare, classify, translate, summarise, interpret, criticise and transform texts without requiring a human reader at every intermediate stage. The familiar question of whether artificial intelligence can "write" therefore conceals a more radical transformation. Increasingly, a human can produce a text without writing it and acquire the informational consequences of reading a text without reading it. The emerging figure is neither Barthes' Author nor his Scriptor but the AI Prompter: a human operator who initiates linguistic processes while potentially performing neither the composition of their outputs nor the reading of their inputs. The article traces this transformation through three stages. First, it returns to Barthes and Foucault to show that the twentieth-century critique of authorship displaced rather than eliminated anthropocentric sovereignty, preserving the Reader as the final human centre of textual meaning. Second, it examines the transition from traditional reading through search and digital retrieval towards delegated machine reading, arguing that generative AI transforms the library from a repository that must be read into an active linguistic environment that can be interrogated. Third, it connects this development to Peter Ayolov's concept of Lingua Citans, understood not as a conscious or literally autonomous language but as a mechanism of linguistic persistence through repetition, quotation, recombination and changing hosts. When machine-generated language becomes input for further machine processing, the classical sequence Author–Text–Reader gives way to recursive linguistic circuits in which human beings may appear principally at the beginning and end of processes whose intermediate textual stages remain unread by any person. The "death of the Reader" therefore does not predict the disappearance of books, reading or human interpretation. It names the disappearance of their structural necessity. After writing liberated words from the physical presence of the speaker, and Barthes liberated the text from the sovereignty of the Author, generative AI begins to liberate textual processing from the compulsory presence of the human Reader. The central question of the AI age may consequently be not whether machines can write like humans, but what becomes of human textual culture when language no longer needs to pass through human consciousness at every stage of its circulation.