Assessing the Effectiveness of Generative AI in Undergraduate Academic Writing

Abstract

Academic writing ability is an essential skill in higher education; however, undergraduates have continued to face challenges in organizing ideas, paraphrasing, and maintaining academic register. However, the rapid development of new tools in the form of Generative Artificial Intelligence, including applications like ChatGPT, has drawn considerable attention regarding their potential in acting as instructional tools for academic writing support. This research explores the effectiveness of Generative AI in developing the academic writing skill of undergraduates in the context of higher education in Sri Lanka. An experimental design was followed for this research, which included 26 information technology and management field students. The writing ability of the students was measured through an analytic rubric that provided an overall score of up to 20. Generative AI support was provided only during the mid-test period, while the pre- and post-tests were done without AI support in order to measure the level of internalization of the subject. The overall scores of the students in the writing task were analyzed through descriptive statistical analysis and gain scores of learning. The outcome of the investigation has shown that Generative AI can be an effective instructional tool for developing academic writing ability in students when provided in an organized and guided manner.

Ähnliche Arbeiten