Pictograms are produced for the purpose of communicating meaning intuitively, quickly, and non-verbally [1], and should be simple and clear. However, they can cause misunderstanding, confusion, and danger due to their shortcomings, such as problems in recognition and comprehension, and their inability to represent complex content [2]. In contrast, "motion pictograms," which incorporate animation, are more intuitive and capable of accurately understanding and supplementing information, and have the potential to strengthen the significance of pictograms that convey semantic content without the need for language, but their high production cost makes it difficult to spread. Therefore, this study uses AI to automatically motion-enable still picture pictograms. We aimed to supply information that can be conveyed to a large number of people by incorporating the advantages of motion pictograms while making them more intuitive and easier to adopt by the general public. we examined the automation of motion-izing still pictograms using AI and verified the effectiveness of using motion pictograms. In the generation of motion pictograms using AI, it became clear that improvements suitable for pictogram expression are needed. It is necessary to consider and develop improvements such as automation of prepro cessing and improvement of the accuracy of 3D human body mesh reconstruction for pictograms by estimating the position coordinates of joints. Regarding the effectiveness of the system, it was found that the system had a certain effect on improving the level of comprehension compared to still images. In addition, we were able to identify necessary conditions, characteristics, and trends regarding the conditions that improve the degree of comprehension and effectiveness. UNSOLVED PROBLEMS The unsolved issues in this research is the adaptation to pictogram representation in the generation of motion pictograms. Currently, two approaches to solving this issue are attempted as follows. ・A method to construct a skeletal estimation system adapted to a deformed human body. ・A method that recognizes the meaning of a still pictogram based on the text and creates a realistic motion based on the text, which is then applied to the pictogram again.