Enhancing the Writing Proficiency of Non-English Major Students Using Artificial Intelligence (AI) Language Tools
Michael Garcia
Received: 27 April 2026; Revised: 16 May 2026; Accepted: 04 June 2026; Published: 09 June 2026
DOI: https://doi.org/10.66074/NET55E77R6
Abstract
The rapid adoption of Artificial Intelligence language tools has fundamentally transformed educational practices, creating a critical need to evaluate their specific impact on the higher-order writing skills of non-English-major college students. This study examined the writing proficiency reflected in the written outputs of non-English major students before and after AI-assisted revision across five distinct dimensions and tested for significant differences to propose a targeted instructional intervention program. Utilizing a quasi-experimental single-group pretest-posttest design, researchers evaluated a purposively selected sample of seventy-seven first-year students enrolled in a general education course. Participants composed an initial unassisted academic essay and a subsequently revised essay guided by Artificial Intelligence feedback, which were then scored using an expert-validated analytic rubric and analyzed using repeated measures ANOVA. The AI-assisted revision intervention was associated with statistically significant improvements in the quality of respondents’ written outputs, elevating the overall mean score from a baseline of 25.55 to 29.97 with a substantial effect size. Despite these gains, originality and creativity, critical thinking, and tone and style remained at the emerging proficiency level, along with contextual understanding, which achieved a proficient mean but still saw the majority of the population at emerging proficiency. However, the observed improvements reflect short-term performance within AI-assisted revision tasks and should not be interpreted as conclusive evidence of permanent independent writing mastery. These findings suggest that while Artificial Intelligence operates as a highly effective structural scaffold, it cannot automatically cultivate the deep analytical and creative capabilities required for authentic scholarly authorship. To address these observed instructional gaps, the study proposes the theoretically grounded Project A.I.-V.O.I.C.E. framework as a potential intervention model designed to strengthen evaluative judgment and maintain human agency during AI-assisted writing tasks. However, the framework itself requires separate empirical validation in future research contexts.
Keywords: AI language tools, augmented intelligence, non-English major students, quasi-experimental study, writing proficiency assessment
Author Information: Pangasinan State University, School of Advanced Studies, Philippines; [email protected]
Volume 2, Issue 2, June 2026