Speaker
Description
In response to concerns about the over-reliance on generative AI technologies for academic writing development, this study examines the comparative strengths and weaknesses of 12 argumentative essays generated by ChatGPT and 12 essays written by undergraduate students in response to the same assessment task. However, rather than evaluate AI-generated writing solely in terms of correctness or deficiency, this study is designed to investigate how comparative differences can inform key dimensions of writing development for learners.
Abstract section 3: Content/method
Grounded in a constructivist view of learning, this study examines the pedagogical implications of AI-assisted writing through a comparative discourse analysis of argumentative essays generated by ChatGPT and essays written by international undergraduate students in English in a Japanese university context. Using computational linguistic methods, we analyzed lexical diversity, syntactic complexity, discourse marker use, and grammar error rates across 24 essays written in response to the same assessment task.
Abstract section 5: References
van Niekerk, J., Delport, P. M. J., & Sutherland, I. (2025). Addressing the use of generative AI in academic writing. Computers and Education: Artificial Intelligence, 8, 100342.
Zahari, M., Marini, A., Safitri, D., Dewiyani, L., & Muawanah, U. (2025). Optimizing student writing performance in higher education: A quantitative study of teacher feedback and classroom environment, Social Sciences & Humanities Open, 101286.
Abstract section 4: Outcomes/results
While AI-generated essays demonstrated strong grammatical accuracy and surface-level coherence, student essays exhibited greater lexical diversity, sentence complexity, and the use of temporal and additive discourse markers. Qualitative analysis further revealed that student writing projected a stronger sense of authorial identity and voice, which was largely absent from AI-generated texts. These findings suggest that although AI tools may be useful for idea generation and proofreading, over-reliance risks undermining the development of critical reasoning, disciplinary engagement, and scholarly voice. This presentation discusses pedagogical approaches that integrate AI cautiously while prioritizing writing as a core process of cognitive development in higher education.
Abstract section 2: Contribution/research questions
What are the comparative weaknesses in the Chat GPT essays that can inform ESL writing teachers and dissuade students from the inappropriate use of AI tools in support of engaging them more in the learning processes to develop their writing skills?
What are the strengths in the AI-generated essays that can guide teachers and students towards a balanced use of technology, whereby AI tools support, rather than replace, the cognitive skills necessary for ESL writing?
Abstract section 1: Relevance
As a teacher having to deal with the misuse of AI tools for written assessment in higher education, the author of this study set out to investigate empirical research findings that can inform teaching and learning practices that develop the cognitive abilities to succeed in ESL writing by comparing the strengths and weaknesses of essays generated by ChatGPT and written by a cohort of undergraduate students in response to the same assessment task.
| Title | Importance of teachers engaging students in writing |
|---|---|
| Teaching Context | College and university education |