Rhetoric, Creativity, and Machine Limits: ChatGPT 4.5 Plus as a Model for Comparing Human and AI-Generated Arabic Prose

While recent advances in generative AI have significantly enhanced the fluency and coherence of machine-produced texts, questions remain regarding the extent to which such systems can emulate the creative, rhetorical, and cultural intricacies of human-authored Arabic prose. This study investigates this issue by focusing on a single advanced language model—ChatGPT 4.5 Plus—as a representative case. It juxtaposes AI-generated prose with selected works by Abdullah İbnü’l-Mukaffa‘, Mustafa Sâdık er-Râfîî, and Ğassân Kenefânî, examining stylistic and rhetorical features through a dual methodological lens: computational linguistic metrics and critical literary analysis. The findings suggest that although ChatGPT 4.5 Plus demonstrates commendable syntactic fluency and textual organization, it currently falls short in reproducing the semantic density, cultural embeddedness, and rhetorical subtlety that characterize human creativity. However, this limitation is understood not as a definitive judgment on AI’s potential, but rather as a reflection of the present capabilities of the studied model. The study concludes by highlighting the enduring significance of human experience in Arabic literary expression and advocates for the development of culturally responsive AI systems, trained on diverse Arabic corpora, to better complement human literary production in the future.