Pola Kebahasaan AI Generatif dalam Karya Ilmiah Mahasiswa: Repetisi Sintaksis, Rigiditas Penanda Wacana, dan Redundansi Semantis: Linguistic Patterns of Generative AI in Student Academic Papers: Syntactic Repetition, Discourse Marker Rigidity, and Semantic Redundancy

Habiburrahman, Habiburrahman
Harsono, Harsono

Pena: Jurnal Pendidikan Bahasa dan Sastra
Vol. 16 No. 3 (2026): Pena: Jurnal Pendidikan Bahasa dan Sastra

The emergence of generative AI tools like ChatGPT has transformed academic writing. At the Indonesian Language Education Program at Madura University, this has led to issues of linguistic rigidity and the homogenization of students’ writing styles. This study aims to deconstruct AI-generated texts to identify the microstructural elements that cause the texts to feel artificial and monotonous. Using a descriptive qualitative approach, functional syntactic analysis, and critical discourse analysis, this study examines a corpus of student papers with an AI probability of >50%. The results reveal three findings: (1) the dominance of repetitive S-P-O-K sentence patterns without variation; (2) the rigidity of discourse markers with clichéd conjunctions; (3) semantic redundancy without adding substantive arguments. In conclusion, the adoption of generative AI without critical human intervention results in a "mechanization of writing style" that threatens the uniqueness of the author's voice and undermines students' cognitive abilities.