Application of Generative AI to Reduce Carbon Footprint from Document Processing in Educational Institutions: A Case Study of NotebookLM

Abstract

Climate change and greenhouse gas emissions have become critical global concerns, prompting educational institutions to recognize the importance of reducing their carbon footprint through more efficient resource management. This study aimed to investigate (1) the learning achievement of educational personnel after training in the use of NotebookLM, a Generative AI tool, (2) their opinions regarding its effectiveness in reducing document-related workload, and (3) the potential reduction in carbon footprint resulting from decreased usage of paper and electricity. The research employed a quasi-experimental design with purposive sampling, involving 44 participants from higher education institutions. Research instruments included pre- training and post-training achievement tests, a structured opinion questionnaire, and a carbon footprint estimation tool based on reductions in paper usage and computer operating time. The findings revealed that post-training learning achievement scores were significantly higher than pre-training scores at the .05 level. Regarding user opinions, most respondents agreed that NotebookLM helped reduce the time required for document preparation, minimized duplication in information searches, and increased overall work efficiency. In terms of environmental impact, the participants reported an average reduction of 10 sheets of paper and 12 hours of computer usage per person per month. These behavioral changes contributed to an estimated total reduction of 372.29 kilograms of CO₂ emissions per year for the entire group. The study demonstrates that the integration of AI technologies in educational administrative tasks not only enhances work efficiency but also offers tangible benefits in supporting institutional sustainability goals.