Pengembangan E-Modul Deep Learning Berbantuan AI untuk Meningkatkan Kesadaran Sejarah dan Hasil Belajar Menggunakan Model Addie

Abstract

This development research aims to create an AI-assisted Deep Learning E module educational product using the ADDIE development model, produce a Deep Learning E-module validated by experts, and measure the effectiveness of the product on increasing historical awareness and student learning outcomes. This research was conducted at SMAN 3 Jember in class XI 9 as an experiment and XI 10 as a control in November 2025 and April 2026. Data collection was carried out through observation, questionnaires, interviews, tests, and documentation. The development of the E-module product has gone through expert validation stages, individual trials, group trials and large trials. The results of the expert validation test show that, (1) the validation of the study field obtained a value of 4.4 (Above Average) which means the product is suitable for use with minor revisions, (2) the validation test of the material obtained a value of 4.5 (Above Average) which means the product is suitable with minor revisions (3) while the validation of the language obtained a value of 4.6 (Outstanding) which means the product is suitable for use without revision. The results of the individual trial by educators obtained a score of 4.3 (Above Average) and the individual test for students obtained a score of 4.9 (Outstanding). The analysis of small group tests was carried out using Paired Samples T-Test and Effect Size Test, then for large group trials using Independent Sample T-Test and N-Gain Test. The results of the small group trials showed a significant increase in historical awareness and student learning outcomes because in the Paired Samples T-Test obtained a Sig. 0.002 <0.05. The Effect Size Test value in the small group trial was 0.89 for historical awareness, and for learning outcomes it was 0.88. Both values are included in the criteria for a large effect. Meanwhile, the results of the large group trial showed that the experimental class (XI 9) using the AI-assisted Deep Learning E-module experienced an increase in historical awareness and learning outcomes compared to the control class (XI 10). The results of the Independent T Test in the experimental class (XI 9) showed Sig. 0.000 < 0.05, and the N-Gain Test (historical awareness) was 0.7028 (g > 0.7). Meanwhile, the N-Gain Test (learning outcomes) was 0.7042 (g > 0.7). Based on the effectiveness test results, the development of the AI-assisted Deep Learning e module product was proven to be moderately effective in improving historical awareness and learning outcomes.