AI or human? A linguistic investigation of generative AI writing

Generative artificial intelligence (gen AI) tools such as ChatGPT, Microsoft Copilot, and Google Gemini are changing how we write. More and more students are using gen AI to plan, structure and edit their work. But a significant question remains: can recognizable hallmarks of gen AI be identified in these texts?

This project endeavors to tackle this question through linguistic investigation. Using similar texts produced with and without gen AI assistance, students will observe and analyze how the act of writing is being transformed.

Students will learn to identify concrete markers, such as:

• general or stereotypical formulations;  
• syntactic or lexical repetitions;  
• unusually homogenous structures;  
• fluid style, but impersonal tone or lack of authorial voice.  

Students will act as true “language detectives,” systematically analyzing texts to better understand what gen AI changes—or doesn’t change—about writing practices.

The objectives are to give students the opportunity to:

1. understand how gen AI can transform the act of writing;  
2. learn to identify and describe linguistic and discursive markers;  
3. contribute to the development of a simple analysis grid;  
4. become familiar with the main steps of performing humanities research.  

No prior research experience is required. Key qualities for this project include curiosity, observation and the development of a critical eye to texts, including their own. This project is part of a wider research initiative on how the act of writing is changing under gen AI. This exploratory step aims to test analysis tools and identify preliminary trends. 

The project takes a qualitative, exploratory approach in applied linguistics and discursive analysis. Students will work with a variety of educational documents composed with and without gen AI assistance, compiled based on the project’s needs. Based on real texts, these documents will offer students a controlled way to observe the different phenomena associated with assisted writing.

A brief guided comparative analysis of scientific and popular texts (2–3 documents) will kick off the project to position students’ analysis within a wider context and to familiarize them with certain issues in gen AI and university writing. This activity, guided by specific questions, is intended to enable students to critically grasp the main ideas rather than to provide an exhaustive analysis.

A structured, accessible analysis grid will be provided to encourage students to focus on:

• style (formulations, repetitions, tone);  
• organization of the text (structure, progression of information);  
• objectivity markers and erasure of voice;  
• markers of the author’s position.  

The course will have five components:

1. Initial orientation: presentation of key concepts and familiarization with tools;  
2. Guided critical analysis: reading short texts and answering targeted questions;  
3. Guided textual analysis: using the grid to identify phenomena;  
4. Data organization: compiling observations;  
5. Analysis and interpretation: identifying trends.  

During this project, students will:

• develop linguistic and discursive textual analysis skills;
• observe, categorize and interpret linguistic and textual phenomena;  
• build a foundation in reading scientific and popular texts critically;  
• better understand university writing norms;  
• take a critical view to the use of artificial intelligence;  
• experience their first foray into humanities research.