Studio Ghibli and other Japanese publishers want OpenAI to stop coaching their work

A Japanese trade organization representing publishers like Studio Ghibli wrote a letter to OpenAI last week, calling on the AI ​​giant to stop training its AI models on copyrighted content without permission.

Studio Ghibli, the animation studio that produced films like “Spirited Away” and “My Neighbor Totoro,” has been particularly impressed by OpenAI’s generative AI products. When ChatGPT’s native photo creator was released in March, it became a popular trend for users to ask to recreate their selfies or pet photos in the studio’s movie style. Even OpenAI CEO Sam Altman changed his profile picture on X to a “Ghiblified” photo.

Now, as more people get access to OpenAI’s Sora app and video generator, Japan’s Overseas Content Distribution Association (CODA) has asked OpenAI to refrain from using its members’ content for machine learning without permission.

This demand does not come suddenly. OpenAI’s approach to copyrighted content is to ask for forgiveness, not permission, which has made it very easy for users to create images and videos of copyrighted figures and dead celebrities. This approach has led to complaints from organizations like Nintendo, as well as the estate of Dr. Martin Luther King Jr., who could easily be deepfaked on the Sora app.

It is up to OpenAI to choose whether or not to cooperate with these requests; If not, aggrieved parties could file a lawsuit, although US law remains unclear about the use of copyrighted materials for AI training.

There are few precedents yet to guide judges on their interpretation of copyright law, which has not been updated since 1976. However, a recent ruling by US Federal Judge William Allsup found that Anthropic did not violate the law by training its AI on copyrighted books – yet the company was fined for pirating the books it used for training.

But Japan’s Overseas Content Distribution Association (CODA) claims this may be considered copyright infringement in Japan.

“In cases, such as with Sora 2, where specific copyrighted works are reproduced or similarly generated as output, CODA considers that the process of copying during the machine learning process may constitute copyright infringement,” CODA wrote. “Under Japan’s copyright system, prior permission is generally required to use copyrighted works, and there is no system that allows a person to avoid liability for infringement through subsequent objections.”

Hayao Miyazaki, one of Studio Ghibli’s central creative figures, has not commented directly on the prevalence of AI-generated interpretations of his works. However, when he was shown an AI-generated 3D animation in 2016, he responded that he was “absolutely disgusted”.

“I can’t watch these things and find them interesting,” he said at the time. “I feel strongly that this is an insult to life itself.”