The Content Overseas Distribution Association (CODA), a key organization representing the collective interests of many of Japan’s most influential entertainment and media corporations—including renowned entities such as the television production giant Toei and the globally recognized video game developer Square Enix—has recently issued an extensive written petition directed at OpenAI. In this document, CODA formally demands that the U.S.-based artificial intelligence company cease what it describes as the unauthorized and inappropriate use of Japanese-owned intellectual property in the training of its newly launched generative AI tool, known as Sora 2. The association’s request reflects growing unease among Japanese creators and cultural producers about the unregulated exploitation of creative assets in AI development, particularly when those assets form part of Japan’s globally celebrated cultural heritage.
The letter, co-signed by nearly twenty influential entertainment and technology entities, accuses OpenAI of infringing on Japanese copyright law by allegedly incorporating a substantial volume of copyrighted materials—ranging from visual art and animation frames to digitally illustrated game assets—into the training datasets used for Sora 2. The signatories assert that a significant portion of the model’s outputs bear an unmistakable resemblance to the visual language, stylistic features, and distinct aesthetics of Japanese media, suggesting that protected works were included in the company’s machine‑learning process without appropriate authorization or compensation. CODA’s correspondence explicitly cites OpenAI’s own stated policy, which currently operates on an opt‑out basis, meaning that the company assumes it has permission to use copyrighted works unless the rights holder explicitly requests removal. However, the signatories maintain that this framework runs counter to Japan’s legal requirements, where copyright norms typically demand prior consent—an opt‑in system—before any creative work can be lawfully reproduced, adapted, or utilized for derivative technological purposes.
In direct response to these issues, CODA outlined two central requests. First, it insists that OpenAI must refrain from using any visual, technical, or narrative content belonging to its member organizations in the training or operation of Sora 2 unless explicit, documented permission is granted by the rights holders in advance. Second, the association expects OpenAI to engage transparently and in good faith with inquiries and formal claims from Japanese companies regarding suspected copyright infringements or ethical breaches linked to Sora 2’s generated outputs. These stipulations underscore CODA’s stance that respect for creators’ rights must serve as a foundational principle in the evolution of AI technologies rather than an afterthought introduced only after public controversy arises.
The escalating dispute comes against the backdrop of growing political attention to the protection of Japan’s creative industries. In mid‑October, the Japanese government itself formally urged OpenAI to halt the unlicensed use of locally produced intellectual properties, including beloved franchises such as *One Piece* and *Demon Slayer*. At that time, Minoru Kiuchi, the Minister of State responsible for intellectual property and artificial intelligence strategy, publicly described these cultural works as “irreplaceable treasures” of national significance, highlighting their role as both economic assets and symbols of Japan’s artistic identity. Other government officials echoed his concerns, framing the matter as part of a broader ethical conversation about how advanced machine‑learning models may inadvertently appropriate or mimic the creative output of human artists.
Earlier in the year, OpenAI’s Chief Executive Officer, Sam Altman, had emphasized the model’s impressive capacity to produce “Ghibli‑like” images during a product demonstration tied to a recent ChatGPT update. Although the statement was seemingly intended to illustrate the visual fidelity and stylistic sophistication of the company’s new generative tool, the remark provoked widespread criticism, particularly after some of the resulting imagery was later employed in political media aimed at reinforcing anti‑immigration narratives within the United States. The incident intensified international scrutiny over the ethical and cultural implications of incorporating copyrighted materials from globally cherished studios—such as Studio Ghibli—into datasets used for commercial AI ventures.
As of the latest reports, OpenAI has yet to issue a direct reply to CODA’s demands. Nonetheless, the association and its member companies have released a more comprehensive follow‑up statement warning that they are prepared to take all necessary legal and ethical measures to address and prevent any further violation of their intellectual property rights. This declaration makes clear that, regardless of whether their organizations adopt generative AI technologies for their own creative purposes, they intend to enforce the protection of Japanese cultural output both domestically and on the international stage.
[via Automaton]
For those interested in related developments, you can stay informed about upcoming releases and industry news, including forthcoming installments in the *Marvel*, *Star Wars*, and *Star Trek* franchises, as well as updates about what lies ahead for the *DC Universe* across film and television, and the future direction of *Doctor Who*.
Sourse: https://gizmodo.com/japanese-companies-tell-openai-to-stop-infringing-on-its-ip-2000680459
