Speaker
Description
Abstract
The rapid adoption of artificial intelligence (AI) tools such as ChatGPT in English as a Foreign Language (EFL) writing classrooms has reshaped how students approach writing tasks, with increasing evidence that learners rely on AI for generating and refining written content. While such tools are widely believed to enhance writing quality, concerns have emerged regarding a potential gap between AI use and actual skill development. This study investigates the extent of AI over-reliance and its impact on students’ writing performance. Adopting a mixed-method design, the study combines survey data with comparative pre- and post-writing analysis. Students completed writing tasks before and after a period of AI-assisted practice, and their texts were evaluated using analytic criteria, including grammatical accuracy, lexical diversity, and idea originality. The participants were 100 non-English major students at Ho Chi Minh City University of Technology. Survey findings suggest that nearly four-fifths of participants relied on AI for idea generation, yet pre- and post-writing scores showed no statistically significant gains across all three criteria. These findings suggest that excessive dependence on AI may reduce cognitive engagement and limit the development of independent writing skills, highlighting the need for more structured integration of AI in writing instruction.
Keywords: ai over-reliance, efl writing, ai-assisted learning, learner autonomy, writing development
