Page 77 - RASAS 2025

th
10 Ruhuna Arts Student’s Annual Sessions (RASAS) -2025

AI Poetics

A.B.M.F.‍ෙAshfa
Department of English and Linguistics, University of Ruhuna
azarbanoon9@gmail.com

ABSTRACT

This study investigates the nexus between literary creativity and artificial intelligence through an analysis of
poetry written by Large Language Models (LLMs), particularly GPT-4. With AI systems playing an

increasingly prominent role in the creative space, long-held assumptions about authorship, emotional
complexity, and literary merit are called into question. Based on post-structuralist theories of authorship and
posthumanist theories of distributed creativity, this study examines how AI models imitate poetic structure and

whether readers emotionally respond to such poetry. The overall goals include analyzing the thematic and
stylistic features of AI-created poetry and gauging human reception of it. In terms of methodology, the research

integrates textual analysis with a reader-response experiment with 100 participants. Participants were asked to
rate a specially curated set of poems—some human-written, others GPT-4-written—on emotional impact,
aesthetic merit, and human-like authorship. Preliminary findings showed that most participants (68%) rated

most of the AI poems as emotionally moving and stylistically sound compared to human-written material. Yet,
when participants learned about the non-human origins of the poems, opinions shifted and many judged the

work as less authentic or emotionally richer. This shift highlights the significance of perceived authorship in
literary worth. The study concludes that AI poetry accounts for a posthuman condition of literature, where
creativity is not to be located within a solitary human author but emerges from interactions between algorithms,

data, programmers, and interpretive readers. This rethinking places “AI poetics” as a rightful area of digital
literature that provokes a rethinking of creative agency, authorship, and aesthetic judgement in the age of
intelligent machines. Beyond literary theory, the research has educational, cultural-production, and

argumentative implications for the ethics and authenticity of AI-produced art. The study invites literary critics,
instructors, and artists to work critically with AI alongside human creativity in the new environment of digital
culture.

Keywords: Authorship, Creativity, Digital Literature, Posthumanism, Reader-Response Theory

49