Integrating Project-Based Learning and Artificial Intelligence (PjBL-AI) Model to Enhance Students' Descriptive Writing Skills

Abstract

This study investigates the effect of integrating the Project-Based Learning model with Artificial Intelligence (PjBL-AI) on university students' descriptive writing skills. Using a quasi-experimental pretest–posttest control group design, the research involved third-semester students of the Indonesian Language and Literature Education Department at Universitas Negeri Makassar. Two classes were selected: the experimental group, which received instruction through the PjBL-AI model, and the control group, which was taught using conventional methods. Data were collected through a descriptive writing test and a student perception questionnaire on AI use, and analyzed using normality, homogeneity, Paired Sample t-test, and Independent Sample t-test analyses. The results showed that the experimental group achieved a significant 25.5% improvement in writing scores, compared to an 8.5% insignificant increase in the control group. The PjBL-AI model effectively enhanced students' content development, text organization, vocabulary, grammar, and writing mechanics. AI integration facilitated cognitive and metacognitive processes through real-time feedback, supporting the stages of planning, drafting, revising, and publishing. This study reinforces constructivist and sociocognitive theories in writing instruction, demonstrating that PjBL-AI can serve as an innovative and adaptive pedagogy that promotes digital literacy and higher-order thinking. Overall, this research introduces a novel approach to AI-assisted project-based learning and provides empirical evidence for its effectiveness, offering valuable insights for future studies and curriculum innovation in language education.