Writing against the machine: Postdigital resistance in LLM collaboration.

Abstract

This chapter examines the creative and theoretical tensions that emerge when postdigital fiction authors collaborate with large language models (LLMs) trained on hegemonic discourse and structures. LLMs, designed primarily by and for dominant (cis-het white male, American corporate) perspectives, internalize and reproduce these patriarchal structures in their textual outputs. Their ever-expanding use exemplifies the postdigital, as they are integrated into daily lives and practices from autocompletion helper functions in almost all digital tools to full-on replacement of human programmers, making them ubiquitous and something more than mere “tools”. Yet creative interaction with LLMs presents a paradox for marginalized writers, particularly femme, queer, and non-Western authors, who seek to use these tools not as passive generative aids but as sites of resistance, subversion, and alternative storytelling. Drawing on feminist, postcolonial, and queer theories of authorship, this practice-based study interrogates the extent to which LLMs can be co-opted and subverted to serve postdigital creative aims rather than reinforcing pre-existing hierarchies.