To effectively apply AI technology in the real world, AI models must continuously learn and grow by integrating new information while leveraging past knowledge. My research is Continual Learning and it aims to enable AI models to learn continuously in such environments. Specifically, I focus on the “Replay method”, akin to human note-taking, where a portion of learned data is retained for future learning. I investigate how much data should be retained for effective ongoing learning. Ultimately, I aim to apply this technology for societal benefit, contributing to the overall digitalization of society, including small-scale organizations and communities.