Effect of Al-Assisted Instruction on Iranian EFL Learners' Writing Performance and Motivational Regulation Strategies

This study investigated the impact of AI-assisted instruction on the writing performance and motivational regulation strategies of Iranian English as a Foreign Language (EFL) learners. Employing a quasi-experimental design, ۵۰ learners were assigned to an experimental group (N = ۲۵) receiving AI-assisted writing instruction and a control group (N = ۲۵) receiving traditional instruction. A multivariate analysis of covariance (MANCOVA) was conducted to examine the effects of AI-assisted instruction on post-test writing performance and motivational regulation strategies, with pre-test scores as covariates. Additionally, a multiple regression analysis explored the predictive role of five motivational regulation strategy subscales-Interest Enhancement, Performance Self-Talk, Master Self-Talk, Emotional Control, and Environment Structuring-on writing performance. The results revealed that AI-assisted instruction significantly enhanced writing performance (F (۱, ۴۶) = ۷۳.۱۵۰, p <.۰۰۱, partial n² =.۶۱۴) and motivational regulation strategies (F (۱, ۴۶)=۲۳۰.۱۵۲, p <.۰۰۱, partial n² =.۸۳۳) compared to traditional instruction. No significant interaction was found between AI-assisted instruction and motivational regulation strategies in determining writing performance (p =.۴۲۱). The regression analysis indicated that the motivational regulation subscales collectively accounted for ۲۰.۷% of the variance in writing performance (F (۵, ۴۴) ۲.۲۹۸, p =.۰۶۲), with Master Self-Talk and Interest Enhancement as the strongest predictors, though not statistically significant (p =.۰۶۱ and p.۰۵۹, respectively). Despite violations of normality and homogeneity of variance assumptions due to the ordinal nature of writing performance scores, the robustness of MANCOVA with equal group sizes supports the findings. These results underscore the efficacy of AI-assisted instruction in enhancing EFL learners' writing skills and motivation, offering implications for integrating AI tools into EFL curricula. Limitations, including the non-normal distribution of writing scores and small sample size, suggest avenues for future research, such as longitudinal studies and qualitative explorations of learner experiences with AI tools.

کلیدواژه ها:

Artificial Intelligence ، AI-assisted learning ، L۲ Writing ، Motivational regulation ، EFL learners