After leaving a struggling gaming marketing agency and assessing my priorities, I knew I wanted to build something my fiancée, a huge anime fan, would enjoy. I leveraged my decade of experience in video production and motion design, and combined it with cutting-edge AI software like ChatGPT, Stable Diffusion, Automatic1111, and AnythingV3.
My original goal was to complete the project in a week, but I soon realized that was unrealistic. Instead, I set a new goal of 100 hours, and with help from the A.I, I managed to create over 400 characters, 200 concept images, and 49 backstories, all voiced and edited by myself.
If I had hired artists to do this work at professional day rates, it would have cost anywhere from $64,500 to $283,201 and taken 625 to 1826 days to complete. But with the help of Stable Diffusion and Automatic1111, the only cost was my electricity bill and subscriptions to Adobe, Artlist, and Envato.
While there are other anime art generators on the market, like NovelAI and Holara, none are as free and flexible as Stable Diffusion. However, there is currently a lawsuit against Stable Diffusion with a lot of misinformation about how diffusion actually works. It will be interesting to see how it plays out. It could either force the art world to adapt and output more high-quality work, or it could give corporations more power over copyright, making it difficult for small artists to produce original work without fear of infringement.
Overall, this project has been an incredible learning experience and I’m excited to see where AI-generated/ assisted content goes in the future.
