Jones, Nathan (2025) The Value of Language Breakdown in the Age of Generative AI. In: Speculative Machines and Us: Histories and Futures of AI, 2025-07-17 - 2025-07-17, LICA.
![[thumbnail of lancaster speculative machines pres]](https://eprints.lancs.ac.uk/style/images/fileicons/slideshow.png)
lancaster_speculative_machines_pres.pptx - Other
Available under License Creative Commons Attribution.
Download (18MB)
Abstract
This talk explores the generative failures of AI as a site for literary and artistic inquiry, arguing that breakdowns in language—rather than fluency—offer the most fertile ground for understanding the value of human creativity in the age of large language models. Drawing on theoretical insights from David E. Wellbery, Gillian White, and Beti Marenko, as well as literary examples from Mircea Cărtărescu and artistic experiments by Joey Holder, the presentation examines how language operates not merely as a vehicle of meaning, but as a viral, contingent, and affective medium. Where generative AI aims to smooth and streamline expression, literature and contemporary art generate epistemic friction—disruption, uncertainty, and aesthetic intensity—that opens new perspectives on cognition, authorship, and subjectivity. The argument is made for treating language failures not as deficits, but as productive disturbances that model the unpredictable nature of thought and world-making. In doing so, the presentation invites a re-evaluation of the role of experimental literature, immersive art, and speculative writing as critical environments for confronting the epistemic limits and aesthetic possibilities of machine-generated culture.