The Effectiveness of a ChatGPT-Integrated Project-Based Academic Writing Learning Model in Improving Students’ Writing Skills

Teaching academic writing in higher education requires approaches that support students’ writing processes and outcomes while responding to the challenges of the digital era. This study examined a ChatGPT-AI-integrated project-based learning (PjBL) model for academic writing among the Indonesian Language and Literature Education Study Program (PBSI) students at Universitas Negeri Gorontalo (UNG). A quasi-experimental pre-test–post-test design was employed with two pre-existing classes: an experimental group (n = 30) and a control group (n = 30). Academic writing was assessed using four indicators: goal achievement, coherence and organization, diction, and sentence structure and grammatical accuracy. Data were analyzed using analysis of covariance (ANCOVA) with pre-test scores as a covariate and an independent-samples t-test on gain scores. The experimental group obtained a higher mean post-test score than the control group (81.17 vs. 65.00). The adjusted means were 78.53 and 67.64, respectively, with a difference of 10.89 points. ANCOVA indicated a statistically significant group difference, F(1, 57) = 50.084, p < 0.001, partial η² = 0.468. The experimental group also showed a higher mean gain score than the control group (16.17 vs. 5.67), with a significant difference, t(38.86) = −7.814, p < 0.001. Overall, the findings suggest that ChatGPT-AI-integrated PjBL was associated with higher academic writing performance within this sample and design. ChatGPT provided supplementary support for language revision, organization, and vocabulary development, while students remained responsible for evaluating and revising their work. The findings support structured AI-assisted PjBL as an alternative approach to academic writing instruction with appropriate teacher guidance and attention to academic integrity.