OpenAI Faces Backlash from Studio Ghibli, Bandai Namco Over AI-Generated Anime Videos

OpenAI has received a written letter from Japan's Content Overseas Distribution Association (CODA) asking the company to stop training its artificial intelligence (AI) models on copyrighted content. CODA is an anti-piracy organisation that represents major Japanese publishing companies such as Studio Ghibli, Bandai Namco, Square Enix, and others. Notably, Bandai Namco is behind works such as Dragon Ball, Naruto, One Piece, and Gitnama, while Square Enix has created the Final Fantasy series. Similarly, OpenAI's image model recently went viral for the Ghibli trend.

OpenAI Told Not to Train AI Models on Copyrighted Anime Content

In a newsroom post, CODA requested that OpenAI not train its Sora 2 video generation model on copyrighted content owned by its members. Additionally, it also told the AI giant to respond to the claims and inquiries made by the members regarding “copyright infringement related to Sora 2's outputs.”

The letter mentions that the agency was able to confirm that a large portion of content generated by Sora closely resembles Japanese content, and it has determined that the issue occurred due to such content being part of its training data. “CODA considers that the act of replication during the machine learning process may constitute copyright infringement,” it added.

CODA also called out OpenAI's initial policy of opt-out. At the launch of the Sora app, the company had told rightsholders to opt out if they did not want the AI model to generate their copyrighted work. This has since been changed to an opt-in policy. The letter stated, “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.”

Two requests have been made by the agency. First is that OpenAI immediately stops training its AI models on content owned by its members, and second, that the company responds “sincerely” to the claims and inquiries made by the publishing companies.

Notably, in September, Japan's Minister of State for IP and AI Strategy, Minoru Kiuchi, said in a Cabinet Office press conference that the government had sent a formal request to OpenAI not to engage in copyright infringement.