This study explores students' perceptions and attitudes toward using artificial intelligence (AI) to write and complete homework assignments. The research subjects were 72 randomly selected students from SMA Unggulan RUSHD. Data were collected through questionnaires and analyzed using descriptive quantitative methods for closed-ended questions (utilizing percentages and averages) and thematic qualitative analysis for open-ended questions. The results revealed that 88.9% of students used AI tools such as ChatGPT to complete their assignments. Most students found AI helpful, but satisfaction with AI-generated answers varied, with 63.9% stating they were sometimes satisfied, 26.4% often satisfied, and only 2.8% always satisfied. These findings indicate that while AI supports academic tasks, limitations in accuracy and relevance prevent it from fully meeting students' expectations. Clear school regulations are needed to ensure AI is effectively utilized without compromising academic integrity. This study contributes new insights into how students use AI for daily academic tasks and their perceptions of related policies. It also emphasizes the importance of AI literacy among students and suggests that educational institutions develop policies that not only regulate but also educate students on responsible and ethical AI usage, fostering sustainable learning in the digital age.

Higher education must adapt its educational practices to ICT, which is now developed rapidly. There are many different approaches, and constructivism-based e-module development is an alternative. This research used the R&D method by adapting the ADDIE development model, which aimed to produce constructivism-based e-module products that meet the criteria of being valid, effective, and practical for use in the learning process. The e-module developed was validated using an expert validation questionnaire filled out by two expert validators, with a result of 4.32, which was included in the valid category. The effectiveness of the e-module was obtained by 100% of students achieving learning completeness scores and increasing learning outcomes based on a normalized gain analysis of 0.71 (high category). The results of the e-module practicality test showed positive student responses, where 86.67% of statements were included in the very strong category and 13.33% in the strong category. Based on the research results, it can be concluded that the constructivism-based general biology learning e-module developed is valid, practical, and effective.