Artificial Intelligence As a Personalized Learning Partner: Implementing and Assessing an AI-Driven Adaptive Learning Model in Blended Classrooms

Abstrak

This quantitative study evaluates the effectiveness of an artificial intelligence-driven adaptive learning system integrated into blended classroom environments across Indonesian secondary schools. The research involved 480 students divided into experimental (n=240) and control (n=240) groups across twelve educational institutions. Students in the experimental group utilized an AI-powered learning platform alongside traditional classroom instruction for one academic semester (16 weeks), while the control group received conventional blended learning without AI integration. Pre- and post-intervention assessments, coupled with weekly learning analytics, were analyzed using independent t-tests, ANCOVA, and effect size calculations. Results demonstrated significant improvements in academic achievement (t(478)=8.42, p<0.001, d=0.77), learning engagement metrics (t(478)=6.95, p<0.001, d=0.64), and adaptive skill development (t(478)=7.23, p<0.001, d=0.66) in the experimental group. Additionally, AI-personalized learning pathways yielded substantial time efficiency gains, reducing average completion time by 23.5% while maintaining higher comprehension levels. These findings substantiate that AI-driven adaptive learning systems represent promising educational innovations that enhance personalization, promote equity in learning experiences, and improve academic outcomes in blended educational settings