Effectiveness of Using AI Chatbots as Interactive Tutors to Improve Personalized Learning and Learning Outcomes of Eighth-Grade Informatics Students

Abstract

This study was motivated by the need for adaptive and personalized learning support for eighth-grade Informatics students at SMPN 1 Pasaman. This study aimed to examine the effectiveness of an AI chatbot as an interactive tutor in improving personalized learning and Informatics learning outcomes, both partially and simultaneously. A quantitative approach with a quasi-experimental method and a Non-Equivalent Control Group Design was employed. The sample consisted of 63 eighth-grade students, with 31 students in the experimental class receiving AI chatbot integration and 32 students in the control class receiving conventional CBL-based instruction. Data were collected through questionnaires and cognitive tests and analyzed using Multivariate Analysis of Covariance (MANCOVA) to control for pretest scores. The results showed significant partial effects on personalized learning (F = 47.879; p = 0.000), with the experimental group achieving a posttest mean of 66.03, an increase of 19.45 points. A significant partial effect was also found on cognitive learning outcomes (F = 6.248; p = 0.015), with a posttest mean of 81.29, an increase of 27.10 points. Simultaneously, AI chatbot use significantly affected both variables (Wilks’ Lambda = 0.476; F = 31.967; p = 0.000), with an effect contribution of 52.42%. These findings indicate that AI chatbots can serve as interactive tutors that provide adaptive cognitive scaffolding according to individual learning needs