The occurrence of drowning accidents on the beaches is mainly caused by rip currents. In this study, we created an AI model that can detect rip currents in Yuigahama beach. As a result of creating the AI model under various conditions, the accuracy, the precision and the recall rates of the rip current detection were 88 %, 17 % and 100 %, respectively at the final point. As a result of actually operating this AI model on the study beach during summer season, except for stormy wave conditions caused by typhoons, the number of detected rip currents tended in the center of the pocket beach to increase with the wave height and wave energy flux. Also, the number of detected rip currents was 10 % or less per hour in most cases. In these cases, because the wave height and the wave energy flux were relatively small, it was considered that flash rip currents occur on the study beach even under calm wave conditions.