Japan’s creative industry is drawing a line in the sand against AI giants. The Content Overseas Distribution Association (CODA), representing leading publishers including Studio Ghibli, has formally urged OpenAI to stop using copyrighted material in its model training without permission.
The move follows growing concern that AI models are learning from and replicating the unique styles of iconic creators without authorisation or compensation.
The Clash Between AI Innovation and Creative Ownership
Studio Ghibli, known for timeless masterpieces like Spirited Away and My Neighbour Totoro, has been particularly affected by the AI art boom. Since the release of ChatGPT’s image generator earlier this year, social media has been flooded with “Ghibli-style” re-creations of user photos, pets, and even celebrities.
Even Sam Altman, OpenAI’s CEO, joined the trend, briefly switching his X (formerly Twitter) profile picture to an AI-generated “Ghibli” version of himself.
But as OpenAI’s Sora, its text-to-video generator, expands to more users, Japanese creators are becoming increasingly vocal. CODA’s recent letter urges the company to cease training AI systems on Japanese creative content without obtaining explicit consent.
A Global Copyright Grey Area
OpenAI’s “train first, ask later” approach has already sparked tension worldwide. The company faces criticism for how easily its products can recreate copyrighted characters or even generate deepfakes of public figures.
In the U.S., copyright law offers little clarity. A recent court ruling found that Anthropic, another AI firm, didn’t violate copyright law by training on protected books, though it was fined for how it obtained those files.
Japan, however, may interpret the law differently. CODA argues that replicating or generating images resembling copyrighted works during AI training could constitute direct infringement under Japanese law, which requires the permission of the creators before use.
While Hayao Miyazaki, Ghibli’s legendary co-founder, hasn’t commented on this latest wave of AI-generated art, his stance on automation is well-known. When shown an AI-produced animation back in 2016, he famously called it “an insult to life itself.”
His words now echo louder than ever as AI-generated art continues to blur the line between homage and imitation.
