Project-based learning as a pathway to writing autonomy and responsible AI use in university EFL learners

Abstract

Introduction The rapid adoption of artificial intelligence (AI) tools such as ChatGPT for academic writing has raised pedagogical concerns regarding learners' writing autonomy, self-regulation, and ethical use of AI in higher education. This study investigates changes in writing performance and reported AI usage patterns following the implementation of Project-Based Learning (PBL) among university-level EFL learners. Methods A quasi-experimental one-group pretest -posttest design was implemented with A2-level learners from Universidad Central del Ecuador. Participants completed writing tasks modeled after the Cambridge A2 Key exam to assess organization, grammar, vocabulary, and coherence before and after a short-term, one-week PBL intervention conducted under restricted AI conditions. Surveys measured learners' writing processes, self-regulation behaviors, and self-reported ChatGPT usage, while instructor surveys captured perceptions of collaboration and motivation. Results Results indicated a statistically significant improvement in writing performance following the brief intervention [t(43) = 4.15, p < .001], representing a change in paired sample scores between the pretest and posttest. Survey responses indicated frequent use of AI at the beginning of the study and more selective interaction during the intervention period; however, these patterns were based on self-reported data. Teachers described learners as highly motivated and collaborative during PBL activities, although challenges remained in task division and group problem-solving. Discussion Considering the quasi-experimental design, the short duration of the intervention, and the absence of a control group, the results should be interpreted with caution and do not constitute definitive evidence of the causal effect of PBL. Nevertheless, the findings suggest the potential impact of instructional design on writing development and the integration of AI tools in EFL teaching contexts, highlighting the need for more comprehensive and controlled research.

Ähnliche Arbeiten