A discourse analysis of cohesive devices in human and AI-produced personal statements

Abstract

This research paper investigates the significant differences and similarities in cohesion devices between personal statements created by AI and those written by humans for Masters of TESOL applications. Using qualitative methods, this study applies Halliday and Hasan's Cohesion Model to analyse six personal statements: three created by AI and three created by humans. The focus is on reference, substitution, conjunction, and lexical cohesion. The findings revealed that the human-generated texts used richer and more varied cohesion devices, including personal and demonstrative references, substitutions, and diverse conjunctions, which enhanced readability and engagement. In contrast, while efficient and coherent, AI-generated texts often lack personal touch and variety, resulting in a more segmented narrative. In addition, AI-generated texts exhibited a higher level of text difficulty and could be detected by AI tools at a rate of 8.05%, compared to 0% for texts written by humans. This study highlights the need for balanced, cohesive elements in AI-generated texts to ensure authenticity and readability, emphasizing the importance of sophisticated detection tools to distinguish between human writing and AI writing. These insights contribute to the EAP understanding of human–AI collaboration in academic writing and inform writing pedagogy and curriculum design by highlighting ways to incorporate AI awareness and critical engagement into instructional practices.

Ähnliche Arbeiten