The Human Motion Video AI Dataset is designed to capture pose information through posture estimation of individuals in video footage obtained from cameras. While extensive research on human motion recognition has been conducted using video data instead of sensor data, publicly available datasets in South Korea frequently rely on datasets from abroad, such as Human3.6M. However, there is a growing trend of enhancing and utilizing specialized datasets created domestically to meet the unique requirements of specific AI projects. Human motion recognition technology is increasingly applied across the medical, sports, and fitness sectors. This technology is particularly valuable for analyzing the movements of patients undergoing rehabilitation and guiding posture correction through motion analysis to ensure accurate positioning. Research in these applications is advancing rapidly, contributing to the broader impact of human motion analysis in various fields.