Japanese cartoonists slam OpenAI, halting unauthorized use of AI

Japanese comic book creators take issue with OpenAI. The Content Overseas Distribution Association (CODA), which represents Japanese anime and comic book publishers such as Aniplex, Studio Ghibli, Square Enix, and Shueisha, has sent a letter to OpenAI regarding its Sora 2 artificial intelligence model, which it uses to generate videos. In the letter, the association urges the company to refrain from using its members' copyrighted content to train AI without authorization.

According to the letter, published in several American media outlets, the cartoonists argue that many of Sora 2's results resemble specific Japanese works, speculating that they were created based on specific learning data, an act that the Content Overseas Distribution Association says could constitute copyright infringement.

"Under the Japanese system, prior authorization is required for the use of copyrighted works," the letter reads, "and there is no way to avoid such liability." The association also challenges OpenAI's opt-out system, emphasizing that a post-clearance objection does not relieve the association of liability for infringement that has already occurred.

In addition to calling for a halt to unauthorized use, the organization urges OpenAI to respond transparently and truthfully to member companies' inquiries regarding specific copyright infringements in the results generated by Sora 2, "to ensure both the healthy development of artificial intelligence and the protection of the rights of copyright holders and creators." In March of this year, the "Ghibli-mania" exploded online, involving the use of ChatGpt to create Japanese anime-style photos from real images.