DESIGN AND IMPLEMENTATION OF AN AI TEACHING ASSISTANT FOR ENHANCING STUDENTS' SELF-STUDY IN 1OTH GRADE BIOLOGY

DOI: 10.18173/2354-1075.2025-0139

Abstract

While AI applications in education are expanding, a baseline survey revealed that 60% of students (N = 184) expressed concerns that AI might impair their critical thinking. To address this, a six-component self-directed learning (SDL) competency model was employed to evaluate students’ SDL competence before and after learning with the AI teaching assistant (N = 223), with the most notable gains observed in “Inquiry and Questioning skills" (ΔM = 0.57) and “Strategic Study Plan” (ΔM = 0.38). Additionally, a customized AI teaching assistant for Grade 10 Biology was developed through a systematic five-step process to ensure strict alignment with the curriculum and academic accuracy. The tool was implemented in a high school setting utilizing the Human-AI-Human model combined with the Gradual Release of Responsibility (GRR) framework, transitioning students from teacher-led instruction to independent mastery. The pedagogical experiment results indicated significant improvements across all six SDL components, with the experimental class achieving a substantially higher midterm mean score (7.84) compared to the control class (6.97) with a large effect size (d = 0.86). These results reveal that a customized AI teaching assistant, especially when integrated within a structured pedagogical framework, acts as an effective learning scaffold that supports both SDL competence and academic performance.