Development of a Differentiated Physics E-Module Based on Deep Learning Using AnyFlip and an AI Assistant for Grade X Secondary School Students

Abstrak

This study addresses the persistent problem of low science literacy and unfavourable student perceptions of physics, which are often reinforced by uniform, teacher-centred instructional practices that inadequately accommodate learner diversity. The study aimed to develop a differentiated physics e-module grounded in deep learning principles using AnyFlip and an AI Assistant, and to examine its validity, practicality, and effectiveness in improving students’ learning outcomes. A research and development approach based on the 4D model was employed. Data were collected through expert validation, achievement tests assessing conceptual understanding and problem-solving skills, and teacher and student response questionnaires. The findings indicate that the developed e-module demonstrates high validity (mean score = 4.50; Gregory coefficient = 0.85) and high practicality, as reflected in positive user responses. Students’ learning outcomes improved markedly following implementation, with mean scores increasing from 43.5 to 80 and an N-Gain of 0.65, indicating moderate-to-high effectiveness. Overall, the results confirm that integrating differentiated instruction, deep-learning activities, and AI-based scaffolding offers an effective strategy for enhancing physics learning while fostering students’ motivation and learning autonomy.

Kata kunci

• Deep Learning
• ai assistant
• anyflip
• differentiated
• physics e-module