Japan\’s Entertainment Giants Confront OpenAI Over Sora 2\’s Alleged Copyright Violations

OpenAI’s text-to-video model, Sora 2, faces mounting legal and cultural pressure as a coalition of Japan\’s premier entertainment companies accuses it of large-scale copyright infringement. The group, led by the Content Overseas Distribution Association (CODA), has formally demanded that OpenAI cease using protected works to train its AI system.

The coalition includes industry titans such as Studio Ghibli, Bandai Namco, Square Enix, Toei Animation, Kadokawa, and Aniplex. They contend that a significant volume of Sora 2’s output directly mimics the distinctive visual style and essence of iconic Japanese anime and games, suggesting their copyrighted portfolios were used for training without authorization.

The conflict centers on a fundamental clash between OpenAI’s policies and Japanese copyright law. Under Japan’s strict intellectual property framework, utilizing copyrighted material requires prior approval; there is no provision for retroactive opt-out mechanisms. OpenAI’s current system, which allows creators to request removal of their content only after it has been used for training, is viewed by CODA as a direct violation of these protections.

Adding weight to the industry’s stance, the Japanese government has intervened, labeling many of the affected works \”irreplaceable cultural treasures.\” Officials have publicly urged OpenAI to immediately halt the alleged infringement of Japanese artistic and cultural content. The dispute has intensified alongside a viral wave of AI-generated videos that closely resemble beloved franchises like PokémonSuper MarioOne PieceDragon Ball, and Demon Slayer.

In response to growing criticism, OpenAI CEO Sam Altman has acknowledged the concerns and indicated plans to revise the company’s opt-out approach. The potential shift toward an opt-in model would grant rights holders greater control over whether their content is used for AI training. Nevertheless, this incident underscores a broader, unresolved global debate: Can AI companies legally train models on copyrighted content without explicit permission?

For Japan’s creative powerhouses, the stance is unequivocal: their artistry, characters, and narratives are not free resources for AI development. For OpenAI, the future of Sora 2 in one of the world’s most influential creative markets may now hinge on its willingness to adapt and address these escalating demands for accountability.