In recent years, environmental problems have become more serious, and the importance of promoting pro-environmental behavior is increasing. Also, many Instagram people act in the same way by seeing a photo post that looks good, so ”BAE” may have potential to change their behaviors. The purpose of this study is to devise a method to promote their pro-environmental behavior, focusing on this BAE. Specifically, we will verify the action provocation effect of BAE photographs through experiments and case studies. In order to objectively check whether the photograph used in the experiments and case studies is a BAE photograph or non-BAE photograph, a BAE classifier is first developed by using deep learning where 3,242 photographs were used for deep learning and its evaluation. As the result of learning, the average of the correct answer rate that BAE photographs or non-BAE photographs were correctly judged by the learned classifier exceeded about 90% by adjusting threshold levels, so the performance of the BAE classifier was enough to be utilized in the following experiments and case studies.