A coalition of 18 Japanese publishers has announced that they will take more decisive action against all forms of copyright infringement in the future. We summarize.
Offensive against AI violations
The trigger for this merger was the release of Sora 2, an AI tool for automatic video creation developed by OpenAI, in October 2025. The application uses copyrighted works for machine learning, which means that the generated results are said to closely resemble the original content.
The publishers involved are Akita Shoten, Ichijinsha, Shogakukan, Ohzora Publishing, Nihon Bungeisha, Kadokawa, Coamix, Kodansha, Shonengahosha, Shinchosha, Square Enix, Takeshobo, TO Books, Hakusensha, Futabasha, Houbunsha, LEED Publishing and the Japanese Mangaka Association (Nihon Mangaka Kyokai).
Among other things, they criticize the fact that Sora 2 does not rely on an opt-in procedure, but rather an opt-out system. This means that works can be used automatically unless a company explicitly requests their removal.
Collaboration with CODA
For this reason, it has teamed up with the Content Overseas Distribution Association (CODA), not only to take its own legal action, but also to seek a response at the national level, in particular calling on the Japanese government to enact new laws to ensure the protection of content in the age of artificial intelligence.
It also plans to actively work with all stakeholders, including government authorities and other copyright holders, to jointly address what they see as “systematic copyright infringement” classify.
In principle, however, people are open to artificial intelligence. You support both »Further development of generative AI« as well as the goal, one »To create a society in which more people can experience the joy of creative work« – but not at the expense of the authors.
Rather, we strive to find a balance between technological innovation and the preservation of cultural creations »fair, transparent and permanently sustainable« to create environment.
In this context, the publishers laid down three basic principles that they believe operators of AI systems should adhere to in the future: