Share and Follow
Japan has delivered a stern message to OpenAI, urging the tech giant to refrain from incorporating anime and manga into its AI models. A powerful alliance of publishers — featuring industry giants such as Kadokawa, famous for isekai series like KonoSuba and Re:Zero, along with Kodansha (Attack on Titan, Blue Lock) and Shogakukan (Detective Conan, Frieren) — has released a forceful declaration criticizing the data training methods of Sora 2, OpenAI’s latest text-to-video AI technology. Echoing these sentiments, Shueisha, the publisher of Weekly Shonen Jump, issued a similarly critical statement.
The statements, one from the 17-member publishing coalition and another from Shueisha, highlight a surge in infringing content following Sora 2’s debut. They cited numerous examples where AI-generated videos heavily relied on and resembled existing creative works, with Shueisha specifically pointing out breaches of anime and character copyrights. Both entities criticized the AI’s training model, arguing that the existing “opt-out” approach — where creators must actively request exclusion — is deeply flawed. They claim this practice contravenes Japanese copyright law and the WIPO Copyright Treaty, with Shueisha labeling the issue as a mounting infringement crisis necessitating immediate intervention.
The coalition and Shueisha are pushing for significant reforms in AI training methodologies. They are advocating for a transition from the current “opt-out” framework to an “opt-in” system, mandating AI operators to obtain explicit consent from rights holders before utilizing their works. Moreover, they demand complete transparency regarding the data used for training generative AI and insist on fair compensation for creators. Shueisha underscored these concerns by urging government involvement, stating that “a national-level response, including new legislation, is vital for safeguarding content.”
Japanese Publishers Warn OpenAI Over Generative AI & Sora 2 Copyright Infringement
Both the coalition and Shueisha are demanding fundamental changes to how AI models are trained. Their primary demand is a shift from the current “opt-out” model to an “opt-in” system, which would require AI operators to get explicit permission from the right holders before using their work. They also called for full transparency into what data is being used for training generative AI and insisted that creators receive appropriate compensation. Shueisha’s statement reinforced these concerns, adding a specific call for government intervention, stating that “a response at the national level, including the establishment of new laws, is also essential for content protection.”
The publishers clarified that this is not a rejection of AI itself, stating they welcome technological advancement if used correctly. However, they stressed their responsibility to protect the work and dignity of the creators, arguing that AI’s evolution cannot be built on infringement. Both parties affirmed they would take action against copyright violations, with Shueisha vowing to take “appropriate and strict measures.” Their stated goal is to foster a “fair, transparent, and sustainable creative environment” where both creators and users can operate with peace of mind.
This is not the first time OpenAI has faced backlash for its generative AI practices. In the past, the company’s image-generation model fueled a viral “Ghiblifying” trend, where users created content that was virtually indistinguishable from the iconic art style of Studio Ghibli films like My Neighbor Totoro and Spirited Away. While the trend was popular, it also received significant backlash from artists and fans concerned about copyright infringement. At the time, many expected the notoriously protective Studio Ghibli to speak out against the unauthorized use of its style, but no official statements were released by the company.
While generative AI has sparked serious copyright concerns for all creators, Japan’s stance is not a wholesale rejection of the technology. Shogakukan, for instance, has invested in the startup Orange Inc. to use AI for manga translations, aiming to drastically reduce localization times. Similarly, Toei Animation, the studio behind One Piece and Dragon Ball, is using AI to improve their operational efficiency. The industry has also begun testing the capabilities of generative AI, though not without controversy. In 2023, Netflix Japan and WIT Studio produced the short film The Dog & The Boy, which famously used AI-generated images for its backgrounds, and more recently, an anime titled Twins HinaHima utilized AI assistance for the majority of its cuts.
While the Japanese anime and manga industry is clearly open to embracing the technology, rights holders draw a hard line when it comes to copyright infringement, as evidenced by their aggressive, ongoing legal battles against piracy, where publishers and groups like CODA have worked with international authorities to shut down scores of illegal streaming and manga sites.
