CODA demands AI developers stop generating copies of anime and manga

The Content Overseas Distribution Association (CODA) issued a statement on May 27, 2026, demanding that generative AI developers take measures to prevent the unauthorized reproduction of Japanese works. The organization considers the current situation to constitute copyright infringement and to damage the foundations of creation.

CODA has detected that some services generate images and videos identical or very similar to well-known Japanese works, even when the user does not mention any specific work in their request. It also points out that when the AI is asked about the generated content, it responds with the name of the original work, confirming the connection between the result and the protected material.

Three requests to developers

The statement makes three demands. First, to continuously investigate whether services produce results very similar to existing works and to prevent their generation. Second, if such similarity is confirmed, to stop using the content of CODA member companies without authorization to train models. Third, to respond in good faith to the requests and inquiries of rights holders.

CODA rejects the idea that private use by the user serves to dismiss the problem: it maintains that responsibility lies with the service offering the tool at scale. Regarding Article 30-4 of Japan's copyright law, the organization admits that certain uses not aimed at enjoyment may be permitted, but argues that when learning is reproduced in concrete results, that use in fact coexists with a purpose of enjoyment and may be illegal.

The organization insists that AI should serve to support human creativity, not replace it. It also warns that recent cases of unwanted similarities are hindering the legitimate use of these tools by creators.

Originally published on Anitsu by Laura Martínez. Text reviewed and verified for DonOtaku on November 5, 2026; the original date is preserved.