Does AI-Generated Scientific Text Follow Human Argumentation Patterns? A CARS-Based Comparison of Research Article Introductions
Human linguistics introductions vary more in CARS moves than LLM versions, which grow more rigid when given the definitions.
The paper compares published linguistics article introductions with LLM-generated versions of the same papers using Swales' CARS rhetorical model. Human introductions vary more in which argumentative moves they use and in what order. Generated introductions are more uniform, and providing the CARS definitions makes the models more rigid rather than more flexible.
- Human introductions are more flexible in CARS move choice and order.
- Generated introductions are more uniform across the same papers.
- Giving models the CARS definitions makes their output more rigid.
From the original article
1 min read at arxiv.orgLarge 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.
By Abdelrahman Sadallah · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.