The Effect of Generative AI Tools (Grammarly and ProWritingAid) on Students' Writing Skill

Background: The integration of Generative Artificial Intelligence (GAI) into English language education has transformed writing instruction, yet empirical comparisons of its tools remain limited. Purpose: This study investigates the comparative effect of Grammarly and ProWritingAid on EFL students’ academic writing skills. Method: A quasi-experimental design was employed with three intact groups (n=37): Grammarly (n=13), ProWritingAid (n=11), and electronic dictionary control (n=13). Writing pre-test and post-test data were analyzed using the non-parametric Kruskal-Wallis test due to violations of normality and homogeneity assumptions. Results and Discussion: A significant difference was found among groups (χ²=10.544, p=0.005). ProWritingAid yielded the highest mean rank (24.73), followed by Grammarly (21.81), and electronic dictionary (11.35). Post-hoc Mann-Whitney tests confirmed significant gains for both AI tools over the control (p<0.05), but no significant difference between Grammarly and ProWritingAid (p=0.792). Conclusions and Implications: Both AI tools effectively enhance writing skills, with ProWritingAid showing slight superiority in stylistic and structural feedback.