Japanese publishers demand OpenAI stop using their content without permission. Major entertainment companies, including Studio Ghibli, Square Enix, and Bandai Namco, have formally asked OpenAI to stop training its AI models on their copyrighted works without prior permission.
The Content Overseas Distribution Association (CODA) submitted a written request to OpenAI on October 27, 2025, regarding the operation of its text-to-video generator Sora 2, launched on September 30, 2025.
CODA confirmed that a large portion of Sora 2’s output “closely resembles Japanese content or images”. The organisation determined that this was likely due to OpenAI using Japanese works as machine learning data. CODA believes this process may constitute copyright infringement under Japan’s copyright law.
CODA stated that Japan’s system requires prior permission for using copyrighted material. It also emphasised that there is no system allowing one to avoid liability through later objections or opt-out systems.
CODA made two key requests:
The organisation urged OpenAI to act responsibly to protect creators while allowing AI technology to develop in a “healthy” manner.
CODA represents major Japanese content producers, including Studio Ghibli, Square Enix, Bandai Namco, Kadokawa Corporation, and Cygames. The association was founded in 2002 to combat piracy and promote Japan’s content industries globally.
Notably, Nintendo is not a CODA member, despite having raised similar concerns over AI-generated imitations of its characters.
Sora 2 has faced criticism for generating videos that mimic Japanese animation styles. A viral clip showed OpenAI CEO Sam Altman standing in a field surrounded by Pokémon, jokingly captioned:
“I hope Nintendo doesn’t sue us”.
CODA said such examples highlight how closely Sora 2’s creations resemble Japanese works.
Sam Altman acknowledged the cultural influence of Japanese content and promised to give rightsholders more granular control over Sora 2 outputs, similar to the company’s existing opt-in model.
Altman’s posted on X:
While no lawsuit has been filed yet, CODA hinted at possible legal action if OpenAI fails to comply. Under Japan’s strict copyright laws, OpenAI’s “ask forgiveness, not permission” approach could be seen as infringement.
In the U.S., however, legal precedent remains unclear. A recent ruling by Judge William Alsup found Anthropic did not violate copyright law for training on books, though it was fined for pirating them.
The AI and copyright battles keep intensifying. CODA has made a bold move. This is Japan’s strongest stand yet to protect its creative industries.