Artificial Authorship in Contemporary Arabic Poetry: A Stylistic and Computational Analysis of Human–AI Collaborative Texts
Abstrak
Purpose – This study aims to analyze the stylistic characteristics of contemporary Arabic poetry produced through human–AI collaboration and to identify the linguistic and aesthetic patterns that distinguish it from poetry written entirely by human poets. The study addresses the growing debate on authorship, creativity, and aesthetic authority in the context of AI-assisted literary production. Design/methods/approach – This research employs a mixed-methods approach by integrating Arabic stylistic analysis with computational techniques based on Natural Language Processing (NLP). The corpus consists of contemporary Arabic poems produced through collaboration between poets and generative AI systems, which are compared with poems written by human poets during the same period. The analysis focuses on lexical, syntactic, and semantic features, imagery, metaphorical patterns, textual cohesion, and linguistic complexity using statistical and computational modeling techniques. Findings – The findings show that human–AI collaborative poetry demonstrates high lexical diversity and a strong capacity to generate broad metaphorical associations. However, it also tends to display greater semantic repetition and structural homogeneity than poetry authored solely by humans. Human intervention is essential in strengthening thematic coherence, emotional depth, and pragmatic meaning, while AI functions as a creative catalyst that expands linguistic experimentation and poetic imagination. Research implications/limitations – This study is limited to selected contemporary Arabic poems and specific computational features, so broader corpora and more varied AI systems may produce different findings. Originality/value – This study contributes to Arabic digital literary studies by proposing hybrid authorship as a framework for understanding AI-assisted poetic creativity.
Identifier
- OpenAlex: https://openalex.org/W7169728172
- DOI: https://doi.org/10.67245/locana.v1i1.23