Abstract
‘Sakura Vision’ is a tool specialised for generating images of contemporary Japan. It is based on the GPTs of ChatGPT. Sakura Vision generates images of contemporary Japanese people and modern-day Japan with short, simple prompts without needing to specify detailed elements such as nationalities and the scene. In ChatGPT, prompting ‘Japanese people doing XXX’ often generates images of traditional Japanese people wearing kimonos or in Japanese-style rooms, which are far from reality for modern Japanese people. This issue may be caused by people with poor prompt settings playing with image generation, as it is not a part of their daily lives. Using Sakura Vision, we aim to generate images in a way that is easy for Japanese people to use, ensuring that the visuals align with contemporary Japan. This approach not only makes the tool more practical and relevant but also encourages wider adoption of image generation technology in everyday contexts, enhancing its integration and utilisation as a general tool.
