THE ADDED VALUE OF ACADEMIC WRITING INSTRUCTION IN THE AGE OF LARGE LANGUAGE MODELS: A CRITICAL ANALYSIS

This study investigates the integration of Large Language Models (LLMs) into academic writing education, focusing on a case study involving bachelor's degree students. In an era where artificial intelligence (AI) significantly influences educational methodologies, our research explores the effective utilization of LLMs to enhance students' academic writing skills. The study spans three academic years and involves undergraduates enrolled in an academic writing course. The primary assignment required students to produce a minimum three-page paper, including a literature review on a self-selected topic with high-quality references. We employed an experimental approach by comparing each of the selected students' papers (n=24) with a parallel version generated by LLMs (OpenAI ChatGPT-4 and Jenni), starting from the same title and abstract of the student's papers. A blind review was conducted to evaluate the quality of the student-written and LLM-generated papers using identical criteria. Preliminary findings indicate that in the specific context of literature review and academic writing, (a) LLM-generated papers scored higher across most evaluation areas compared to their human-written counterparts in a twentieth of the time usually allowed to a student, and (b) LLM-generated papers were utterly indistinguishable from those written by students during the review process. The study also addresses challenges such as maintaining academic integrity and fostering critical thinking skills in an AI-assisted learning environment. Additionally, side effects on the role of educators in this context are considered, particularly in disciplines heavily impacted by generative AI regarding educational methodologies, content delivery, and evaluation processes.