The Effect of AI-Based Problem Based Learning Model on Improving the Digital Literacy Understanding of Grade VIII Students in Social Studies Learning at SMPN 2 Surabaya

Abstract

In the contemporary digital era, the learning of Social Sciences (IPS) demands that students cultivate robust digital literacy, encompassing not merely the technical operation of devices, but the critical ability to access, evaluate, synthesize, and ethically utilize digital information. However, conventional pedagogical models, which predominantly rely on teacher-centered lecture methods and rote memorization, pose a significant barrier to fostering these essential 21st-century competencies, often leaving students as passive consumers rather than critical thinkers of digital content. Consequently, this quantitative study aims to rigorously analyze the effect of the Problem Based Learning (PBL) model integrated with Artificial Intelligence (AI) on enhancing the digital literacy comprehension of eighth-grade students enrolled in IPS classes at SMPN 2 Surabaya. Employing a quasi-experimental framework, specifically the Non-Equivalent Control Group Design, this research compares an experimental class (VIII-E) with a control class (VIII-I), each consisting of 32 students. Data were meticulously collected via multiple-choice pretests and posttests designed to measure cognitive gains in digital literacy comprehension. The collected data underwent stringent statistical scrutiny using SPSS version 23.0, including normality and homogeneity tests to validate parametric assumptions, followed by N-Gain analysis to measure improvement magnitude, and Independent Sample T-Tests to detect significant differences between groups. The findings reveal that both groups initially demonstrated normal and homogeneous data distributions, satisfying the prerequisites for parametric testing. Crucially, the experimental class achieved a substantial N-Gain score of 0.689, categorizing their improvement as moderate to high, which significantly outperformed the control class that registered a low gain score. Furthermore, the Independent Sample T-Test yielded a highly significant p-value of 0.000 (with a calculated t-value of -11.758), decisively rejecting the null hypothesis. These compelling results conclusively affirm that the AI-enhanced PBL model exerts a significant and positive influence on students' digital literacy. Thus, this innovative pedagogical approach proves highly effective in empowering students to navigate, analyze, synthesize, and apply digital information critically, actively, and contextually within their IPS studies, ultimately preparing them to thrive amidst the complexities of an increasingly digitalized society.