Does AI-Generated Scientific Text Follow Human Argumentation Patterns? A CARS-Based Comparison of Research Article Introductions

2026-10-01

AI Summary

This study investigates the differences in deep rhetorical structure between AI-generated and human-authored scientific texts, addressing the limitation of prior research that has predominantly focused on surface-level lexical features rather than macro-argumentative patterns. Drawing on Swales’ CARS model and computational stylistics, we conduct a comparative analysis of linguistics paper introductions written by large language models (LLMs) and human scholars to examine their rhetorical characteristics and flexibility. Our findings reveal that LLM-generated introductions are significantly more rigid and uniform than human writing, and that providing definitional prompts further exacerbates this homogenization. By moving beyond conventional surface-level analytical paradigms, this work offers new perspectives for evaluating and enhancing the rhetorical competence of LLMs in academic writing.

Abstract

Large language models are moving from helping write up research to helping do it, which makes it important to know how the scientific text they produce differs from human writing. Work on this question has stayed mostly at the surface, using lexical and stylistic cues that light paraphrasing erases. We look instead at rhetorical structure, the sequence of argumentative moves through which a text makes its case. We study research-article introductions under Swales' CARS model, and compare original introductions from published linguistics articles with generated counterparts of the same papers. We find that human-written introductions are more flexible in which moves they use and in what order, while the generated ones are more uniform. Giving the models the CARS definitions makes them more rigid.

Problem

• AI-generated scientific text
• argumentation patterns
• rhetorical structure
• CARS model
• research article introductions

Innovation

• Large Language Models
• Rhetorical Structure
• CARS Model
• Argumentation Patterns
• Scientific Text Generation