Existing scholarship –from conceptual writing to critical code studies– has long recognized the machine as a potential collaborator. In these frameworks, the human programmer provided the instructions (algorithms) and the content (datasets), yet could still be surprised by unexpected textual outputs. However, the emergence of Large Language Models (LLMs) introduces a new “probabilistic” layer where the human writer is no longer in charge of creating the compositional rules or providing the content. By situating GenAI writing within the lineage of Fluxus, Oulipo, and the cut-up method, among others, this session studies “prompting” as a new way of designing constraints. Panelists argue that the writer’s role has shifted from the composition of sentences to the architectural design of linguistic conditions operating within a complicated ecosystem of professional writers, for-profit technologies, and the literary market. This session bridges the gap between 20th-century experimental writing and 21st-century “Ailiteracy” (Saper), investigating how the materiality of prompts shapes both the digital text (Hall) and the physical book (Ortega). Each presenter will speak for 20 minutes, followed by an unstructured conversation with the audience moderated by Saum-Pascual (at least 15 minutes).

Craig Saper begins by examining the prompt as a direct evolution of Fluxus scores and Oulipian constraints, understanding prompting as a literary and sociopoetic collaboration. The open constraints of this often one-line artworks are the methods now used to coax narrative, information, and meaning from an AI chat bot. Using Fluxus to inform the creation and prompts, Saper draws the limits of LLMs to produce original works, highlighting the need for human constraints –much like OULIPO poetry depended on rules. In Fluxus, the open constraints did not play the role of the LLM, but instead the role of a prompt asking for a performance without a closed content. Saper’s prompt experiments reflect their own conditions of production, self-referencing the human-AI collaboration that constitutes their authorship. This practice suggests a new skill, “Ailectracy,” that we must teach –not just repeating Fluxus, but creating a new FluxAIs movement.

Kimberly Hall continues the discussion exploring the threat to creativity and authorship that practices described by Saper are thought to pose. Like Saper, Hall argues that prompting constitutes a contemporary form of procedural authorship in which writers design systems that generate texts rather than composing sentences directly. Complementing the lineage of Oulipo and Fluxus, Hall presents 20th-century experiments such as automatic writing and the cut-up techniques of William S. Burroughs. Later works such as Tom Phillips’ A Humument, and M. NourbeSe Philip’s archive-based Zong!, foreground the transformation of textual archives into experimental literary forms. More recent works like 1 the Road by Ross Goodwin demonstrate how authorship now includes creating systems of custom datasets and hardware. In this context, GenAI prompting functions as an architecture of possibility in which the weighted constraints of genre, narrative structure, and accumulated personal style guide the output.

Élika Ortega shifts the focus to the materialization of this output in print, investigating how GenAI prompts interact with the tradition of artist’s books. GenAI texts have become commodity objects through self-publishing and print on demand platforms, illustrated by Javier del Puerto’s Payload (2025), a collection of short stories published via Kwalia Books which emphatically declares “this book exists.” Although del Puerto’s declaration is a pragmatic one, it brings into focus the distinction between text and book as object. Two projects of digital literature, Shasha Stiles’s Biblio-Tech (2023-2024) and Richard Carter’s These Books Do Not Exist (2025) address this problem of materialization. Stiles’s Biblio-Tech is a “fantasy library” of synthetic rare books with impossible structural and material features that construct a sense of rarity. These become objects and commodities via NFT minting, their materiality now entangled in the well-known costly process of energy consumption and environmental impact. In contrast, Carter’s These Books Do Not Exist is an extensive collection of AI generated images of page spreads, showing a variety of artbooks and other publications distributed via Creative Commons. Unlike Stiles, These Books Do Not Exist emphasizes the non-unique and non commercial nature of any of these creations. Ortega examines how these two projects reveal the materialization and circulation of prompting.