EVALUATING THE IMPACT OF AI-ASSISTED TOOLS ON ACADEMIC WRITING SKILLS AMONG UNIVERSITY STUDENTS IN PAKISTAN

Abstract

This quantitative research reveals the possible effects of AI technologies on students' academic writings, writing self-efficacy, and responsible technology use in the higher education system in Pakistan. It also explores the possible impact of AI writing assistance on the four main aspects of writing, plus students' self-confidence, and it is based on Bandura's Social Cognitive Theory of Self-Efficacy and Technology Acceptance Model (TAM). A sample of 200 undergraduate and postgraduate students from various universities in Pakistan was collected using three standardized instruments: the Writing Self-Efficacy Scale (WSES), the AI Attitude and Responsible Use Scale (adapted from the Technology Acceptance Model or TAM), and the Academic Writing Skills Rubric. Results indicated that frequency of AI use stood positively correlated to writing self-efficacy (r = .35, p < .01) and academic writing performance (r = .32, p < .01). Results of multiple regression tested were shown that frequency of AI use stood positively in relationship to writing performance (β = .32; p < .001) and accounted for a further 10% of the variance in writing outcome (R² = .10). It was noted at the outset of the study that relentless AI users achieved better performance than moderate users in grammar-by one-way ANOVA, F= 6.20; p < 0.001-and vocabulary (F= 5.45; p=0.002); interestingly, however, moderate users scored significantly higher in the area of cohesion (F= 4.80; p=0.003), thereby supporting the hypothesis that optimal writing may arise from a balance of both human and AI intervention. Gender differences regarding attitudes toward AI and its outcomes were not statistically significant (p = 0.521). The findings support the proposition that if properly utilized, AI could be a valuable tool to scaffold academic writing and writing self-efficacy. The study uses the critical engagement of AI tools for academic writing, digital literacy training, and ethical guidelines to maximize the use of AI for academic writing without compromising originality and critical thinking. It also discusses limitations and recommendations for future research, including longitudinal and cross-cultural studies.