The Novel in the Age of AI (Day 1)

A workshop exploring how AI shapes novels, literary creativity, criticism, publishing, audiences, and the role of literature in training AI Program Thursday, November 5 (Day 1) Kevin McLaughlin: Welcome Remarks Session 1 (3:00-5:30 p.m.) Kevin McLaughlin: “Entailment and the French Novel: Proust on Flaubert’s Style” A key question, and indeed what is generally perceived as a vital threat, that is being posed by what the computer scientist call “textual entailment” is the possible effacement of the distinction between the automated “entailment” and the human “inference.” This presentation takes up this question in the context of Proust’s famous essay on Flaubert’s “style.” In Proustian terms the question is: at what point does the mechanical or involuntary imitation of the Flaubertian text turn into one that is human and voluntary? For Proust, it is the point at which the allowing of the involuntary imitation becomes spontaneous and the involuntary and the voluntary interpenetrate. The linguistic system of “a Flaubert” that is involuntarily imitated must be able to generate a text that, when allowed, can become voluntary in a way that does not entirely exclude the involuntary and that makes it impossible to distinguish absolutely between the one and the other. Nan Da: “Two Kinds of Frightening Happenings in Henry James’s Fiction” This paper is interested in two kinds of frightening things in the literary works of Henry James and current AI’s ability to observe them given specific prompting. They are 1.) a class-based disadvantage that gets one into serious trouble, including wrongful death, abuse, and unassailable maligning by other characters and readers; 2.) the act of truly sneaking up on someone— both how it happens and the way it deranges literary representation and comprehension. This paper will give four ai models (and a smaller, more tailored language model) examples of the two phenomena and ask them to find similar examples across all of Henry James’ literary works (from “A Tragedy of Error,” the first short story he published, to his last unfinished work, A Sense of the Past). Are they accurate findings? What do they reveal about the difference between human inference and a form of scaled inference that combines human inference with nonhuman inference and resources? Can a different understanding emerge of what’s frightening in literature? Zakir Paul: “The Novel and Counterfeit Life” From early modern automata to science fiction and cyberpunk, novelists have long reflected on the porous boundaries between the human and the artificial. These themes are no longer confined to genre fiction. While recent theories of AI increasingly question the distinction between “natural” and “artificial” intelligence (David Bates, Catherine Malabou), theories of the novel reimagine the genre as a prosthetic form of “artificial life” (Peter Boxall). Although LLMs raise seemingly intractable questions for language as such, the production, consumption, and interpretation of long-form narrative fiction may pose a distinct set of problems. At stake is the shifting relation between form and information, fact and fiction, counterfeit and authenticity, personhood and embodiment, as well as the collapsing analogy between experience and narrative. Focusing on modernist debates about intelligence and artifice, this paper argues that current developments in artificial authorship do not pose a fundamentally new problem for the novel but recall one of its constitutive, if continually perplexing, abilities: the capacity to imitate life in its absence. Register for Day 2