AI ENTERTAINMENT
Creators are done letting AI borrow their style
A major Japanese trade group representing studios such as Studio Ghibli has asked OpenAI to stop training AI models on their copyrighted content without permission.
The request follows the rise of “Ghibli-style” images and videos made using OpenAI tools.
When ChatGPT’s image generator launched, it became common for users to create artwork in Ghibli’s style, even Sam Altman briefly used a “Ghiblified” profile photo.
With OpenAI’s new video tool Sora, concerns have grown further.
Japan’s Content Overseas Distribution Association (CODA) says material from its members should not be used for training unless permission is granted.
The group argues that, under Japanese law, copying protected works during training could count as infringement.
Unlike the US, Japan does not allow a “use it first, object later” approach.
Legal clarity remains limited. In the US, there is no settled position on whether training AI on copyrighted works is allowed.
In short
A Japanese trade group has asked OpenAI to stop training on its members’ work without permission.
Japan’s copyright rules require approval before use; there is no opt-out system.
Legal standards around AI training remain unclear globally.
Sora stirred the pot
A recent ruling found that AI training on copyrighted books did not violate US law, though the company involved was penalised for how it sourced the material.
Studio Ghibli co-founder Hayao Miyazaki has not commented on this current dispute.
However, in 2016 he criticised AI-generated animation, calling it “an insult to life itself.”
I think we can all collectively agree to leave Ghibli alone. Or else. - MV