FROM STORYTELLERS TO DATA SUBJECTS: POST-HUMAN LANGUAGE AGENCY IN AI-DRIVEN INTERACTIVE NARRATIVES

Abstract

This paper explores how AI-driven interactive narrative platforms redefine traditional concepts like narrative agency, authorship, and language production through a post-human lens. Drawing on post-humanist theory, enactive cognition, and distributed agency frameworks, the study investigates how generative AI systems—particularly large language models (LLMs) deployed in platforms such as AI Dungeon—participate actively in meaning-making rather than functioning as neutral tools. Using a mixed-method synthesis of theoretical literature and empirical studies on AI-based storytelling systems, the research demonstrates that narrative agency in contemporary interactive narratives emerges from entanglements among human users, computational architectures, training datasets, and platform affordances. The analysis foregrounds a critical transformation of shifting users from autonomous storytellers to hybrid participants who become co-authors, collaborators, and data subjects simultaneously, whose creative efforts contribute to algorithmic optimization. Case studies of GPT-based narrative generation highlight how the narrative possibilities and production of emergent stories are enabled and controlled by text generation, genre recognition, and memory architectures through participatory sense-making. The findings challenge anthropocentric models of creativity, refute human-exclusive agency, and establish the distribution of agency by supporting post-humanist arguments. The paper concludes that AI narrative platforms constitute cultural assemblages that fundamentally modify storytelling practices, raising important implications for narratology, digital humanities, AI ethics, and the governance of creative labor.