Latent Worlds: AI and the Novel - Prof. Hannes Bajohr (Seminar, 12pm & Lecture, 6pm)

Lunchtime Seminar - 12pm, HQ 276, Register on Eventbrite to reserve your seat and receive the advance reading.

Evening Lecture - 6pm, HQ 136 - No registration required.

Latent Worlds: AI and the Novel: 

“A world—nothing less—is the theme and postulate of the novel,” German philosopher Hans Blumenberg wrote in 1963. At that same moment, AI research, already emerging from its early optimism, turned to “world models” as a means of stabilizing its brittle systems. Today, these two conceptions of “world”—the literary and the computational—converge in large language models (LLMs), which use their latent spaces not just to generate plausible sentences, but entire narratives, even novels, albeit with still uneven results. In what sense, this lecture asks, are the “worlds” of novels and of AI analogous at all, and what can each illuminate about the other? One suggestes answer is that both novels and LLMs operate within structured networks of relations—assemblages of events, inferences, and expectations—that can yield a form of coherence even when classical causality is weak or absent. Modernist experiments, with their fragmented but still intelligible world-logic, are the most obvious examples. And where early AI-systems like SHRDLU pursued explicit symbolic world models and failed outside narrow domains, contemporary LLMs rely on distributed vector spaces that encode statistical regularities without grounding. My own experiments with a fine-tuned German-language model yielded narratives with stylistic unity but little causal depth. Like certain experimental novels, they evoke meaning through a “weak force” of association rather than strong narrative causality. This talk tries to follow these ideas and aims to resist both overhyping LLMs’ understanding and dismissing them as mere mimicry, thus placing AI-generated fiction, as the meeting points of the two uses of “world,” within a broader theory of modeling and meaning.

Prior to joining Berkeley’s Department of German in 2024, Hannes Bajohr held postdoctoral positions at Berlin’s Leibniz Center for Literary and Cultural Research, the University of Basel, Switzerland, and Zurich’s institute of advanced studies, the Collegium Helveticum. He has published extensively on the impact of digital writing technologies on language and literature, the German philosophical tradition in the 20th century – especially the connection between phenomenology and anthropology – as well as liberal and republican political theory. A particular interest connects him to figures like Hans Blumenberg, Hannah Arendt, Peter Weiss, and the political theorist Judith N. Shklar, six of whose books he has edited and translated into German. Professor Bajohr is not only a theoretician but also a practitioner of digital literature. He has published numerous generative works as a part of writers’ collective 0x0a, including the volume Halbzeug (Berlin: Suhrkamp, 2018), which was translated into English as Blanks (Denver: Counterpath, 2021). In 2023, he published (Berlin, Miami), a novel co-written with a self-trained large language model. Currently, he is working on a project about “post-artificial writing,” the impact of generative AI on literary reading expectations; and on another on “negative anthropology,” a strand of German philosophy that eschews any definition of an “essence of man” but still insists on making the human the main focus of its attention.