Major Japanese entertainment firms are pushing back against OpenAI’s newest AI video creator, which they say has been using their copyrighted content without permission. The Content Overseas Distribution Association, CODA, in representation of some of Japan’s largest names in entertainment, sent a formal letter of request to OpenAI on October 28.
The group, whose members include powerhouses such as Bandai Namco, Square Enix, Studio Ghibli, Toei Animation, and Kadokawa Corporation, wants the company to cease using their Japanese copyrighted content for training Sora 2, its generative AI video tool.
OpenAI took Sora 2 live on October 1, but it wasn’t a week before the eyebrow-raising began. The tool lets users create short video clips through text prompts, and in virtually no time, social media was flooded with AI videos aping some of Japan’s most iconic characters and animation styles. Users were creating content that looked quite similar to franchises like Pokémon, Mario, One Piece, Dragon Ball, and Demon Slayer.
Japanese Rights-Holders Accuse OpenAI of Copyright Infringement Over Sora 2 Training Data
OpenAI CEO Sam Altman seemed to view this as a good thing, writing in his post-launch blog that “we are struck by how deep the connection between users and Japanese content is!” But what Altman saw as enthusiasm, Japanese rights-holders saw as potential copyright infringement.
The statement by CODA made the position clear: “We have confirmed that a large amount of Sora 2’s output closely resembles Japanese content or images.” The association thus drew its conclusion that the similarity could be due to OpenAI having used Japanese intellectual property as training data without permission. They warned this could constitute copyright infringement under Japanese law.
The request from the association has two demands in particular. Firstly, OpenAI is to cease using the works of its members for machine learning purposes without prior permission effective immediately. Secondly, the company needs to “respond sincerely” to copyright-infringement claims and inquiries from CODA members about Sora 2’s outputs.
Some reports indicate that OpenAI did reach out to some studios and talent agencies, about a week before Sora 2 launched, to offer an opt-out opportunity, but it’s not known whether Japanese rights-holders were contacted. More to the point, CODA noted that Japan’s copyright system depends on prior permission, making an opt-out approach irrelevant anyway in terms of saving OpenAI from liability.
The controversy has gone far beyond the private sector. Earlier this month, the Japanese government formally asked OpenAI to cease infringing on Japanese intellectual property in connection with Sora 2-an unprecedented request that makes this a moment in the global conversation about AI and copyright protection.
Japanese Government Threatens AI Investigation Over IP Rights
At one press conference, Minoru Kiuchi, Japan’s Minister of State for Intellectual Property and AI Strategy, referred to anime and manga as “irreplaceable treasures” that symbolize Japan’s pride in culture.
As such, his remarks have shown how serious Japan is about the issue at hand-this is not just about some commercial products but cultural assets that define the nation’s creative identity.
Kiuchi, as well as Digital Minister Masaaki Taira, made it very clear that failure to comply on a voluntary level would see the government invoke provisions under the AI Promotion Act.
Coming into full force in September 2025, this law grants the government powers to investigate AI utilization when intellectual property rights have been breached. While the act does not place direct penalties, the key threat of government investigation adds severe weight to the situation.
Founded in 2002, CODA has been fighting for Japanese content creators. The organization was founded to fight against piracy and for the legal distribution of Japanese entertainment worldwide, but now finds itself at the forefront of a new battle: protecting creative works from unauthorized uses for AI training.
The question now is how OpenAI will respond. It has built its business on the promise of democratizing AI technology, a promise that brings with it responsibilities regarding intellectual property and copyright law.
As AI tools get ever more sophisticated at mimicking existing creative styles, the legal and ethical questions surrounding their training data are only going to get more urgent.
The message is clear for Japanese creators and the government that supports them: their cultural treasures aren’t free training material for tech companies, no matter how impressive the resulting AI might be.
