TITLE:
The Use of Hedging Devices and Engagement Markers in AI-Generated and Human-Written Essays: A Corpus-Based Comparison
AUTHORS:
Naif Almulla
KEYWORDS:
Artificial Intelligence (AI), LLMs, Academic Writing, Essay Writing, Hedging, Engagement, Corpus-Based
JOURNAL NAME:
Open Journal of Modern Linguistics,
Vol.15 No.5,
September
18,
2025
ABSTRACT: AI tools, such as large language model (LLM) chatbots, have made attaining many learning and educational tasks much easier, and probably even more efficient. These easily accessible and useful resources have made it easy for many students to rely on these tools. This reality has raised some concerns among educators regarding possible negative effects on cognitive abilities and ethical considerations that might result from extensive use of such tools. In order to address this issue, some researchers have attempted to see if LLM chatbots have a unique style that educators can identify and distinguish from the human writing style. Contributing to this effort, this corpus-based study aims to examine two characteristics found in academic writing (i.e., “hedging” and “engagement markers”) to see whether there are differences between AI-generated text and human-written text in the use of these two. In order to analyze and compare the two types of text for hedging and engagement markers, two main text collections were compiled, each consisting of about 20,000 words. Taking a mainly descriptive statistics approach to analyze the data, the results showed that AI-generated text used noticeably more hedging words, whereas human-written text used noticeably more engagement markers. These findings corroborate previous research, provide a better conceptualization of the issue, and emphasize the need for further research on the topic.