AI or student writing? Analyzing the situational and linguistic characteristics of undergraduate student writing and AI-generated assignments

Abstract

Since the release of OpenAI's ChatGPT, universities have faced the issue of whether there is still a place for written assignments in higher education. ChatGPT's capacity to mimic various written forms raises questions about the necessity of traditional assessments. Given this background, this study explores to what extent AI-generated assignments can replicate the situational and linguistic features of student-authored assignments. Using a corpus of undergraduate assignments from an English as a Foreign Language (EFL) context, we compare student responses with ChatGPT's outputs. Employing a register approach, we analyze the situational and linguistic characteristics of texts across three different registers—essays, critiques, and personal narratives. Our methodology follows Biber and Conrad's (2019) framework, encompassing situational analysis, linguistic analysis, and functional interpretation. The findings aim to inform writing instructors and EFL teachers about the strengths and limitations of AI tools, enhancing their ability to guide students in integrating these technologies into their writing processes.