Japan’s CODA demands OpenAI remove Japanese content from Sora 2

Japan’s trade association CODA has officially demanded that OpenAI stop using Japanese creative works in its video generation model, Sora 2. The statement was published on the association’s official blog.

CODA represents some of Japan’s largest media companies, including Bandai Namco, Square Enix, Kadokawa, and the legendary Studio Ghibli. The group argues that OpenAI’s current data practices violate Japan’s copyright laws.

The core issue, CODA says, is OpenAI’s “use first, ask later” approach. Under Japanese law, any use of copyrighted content requires prior authorization from the rights holder, not post-factum removal upon request.

The association also stresses that allowing creators to exclude their works after training does not absolve OpenAI of liability for the initial infringement. It is calling for a complete overhaul of how the company handles Japanese intellectual property.

While CODA says it is open to dialogue with OpenAI, it has not ruled out legal action if its demands are ignored. The group insists that Sora 2’s current policy is unacceptable and must be revised immediately.