Can AI Write Like a Scholar? A Comparative Analysis of Human vs. ChatGPT-Generated Research Article Introductions and Literature Reviews

Abstract

The rapid progression of technology and the proliferation of Artificial Intelligence (AI) have increasingly influenced many fields, ranging from information technology through to tourism, education, management, and medicine. Specifically, in the current academic context, where the ‘publish or perish’ imperative forces constant research and publication, AI technologies can assist academics in writing and improve the efficiency and quality of texts (Dwivedi et al., 2023; Nguyen et al., 2024). However, despite the spectrum of capabilities offered by these tools, growing concerns have been raised about the erosion of critical thinking skills and the production of substandard content. While the impact of the use of AI has received scholarly attention in student academic writing, the effects of its utilization by academics in research articles remain largely unexplored. Therefore, the aim of this study is to investigate how GenAI tools, particularly ChatGPT, could impact expert writers’ research papers in their introduction and literature review sections. The present research compares authentic human-authored literature reviews and their counterparts generated by ChatGPT in response to specific prompts. The focus is on the linguistic, discursive, and rhetorical patterns that (dis)similarly characterize a set of human- vs. AI-generated introductions and literature reviews in the disciplinary field of linguistics. A combination of well-consolidated quantitative and qualitative methodological approaches, including corpus linguistics and discourse analysis, is adopted to assess whether and how ChatGPT chatbots emulate human language use in research article introductions and literature reviews.