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

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

Studio Ghibli, the animation studio behind films like “Spirited Away” and “My Neighbor Totoro,” has been especially impacted by OpenAI’s generative AI products. When ChatGPT’s native image generator was released in March, it became a popular trend for users to prompt for recreations of their selfies or pet pictures in the style of the studio’s films. Even OpenAI CEO Sam Altman changed his profile picture on X to a “Ghiblified” picture.

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

This request does not come unprompted. OpenAI’s approach to working with copyrighted content is to ask forgiveness, not permission, which has made it all too easy for users to generate photos and videos of copyrighted characters and deceased celebrities. This approach has yielded complaints from institutions like Nintendo, as well as the estate of Dr. Martin Luther King, Jr., who could very easily be deepfaked on the Sora app.

It’s up to OpenAI to choose whether or not to cooperate with these requests; if not, the aggrieved parties can file a lawsuit, though United States law remains unclear about the use of copyrighted material for AI training.

There is little precedent thus far 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 Alsup found that Anthropic did not violate the law by training its AI on copyrighted books — the company did get fined for pirating the books it used for training, though.

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

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

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

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