Japan has issued a formal warning to OpenAI, the maker of Sora 2, asking the company to stop generating videos that infringe on anime and gaming copyrights. The government said it may take legal action if OpenAI continues to use characters from Japanese franchises without permission.
The warning comes after Sora 2’s recent release led to a flood of social media clips showing characters like Mario, Pikachu, and other well-known icons appearing in AI-generated scenes. Japan’s Cabinet Office and Minister of State for IP and AI Strategy, Minoru Kiuchi, said such misuse threatens the nation’s cultural heritage. Kiuchi described anime and manga as “irreplaceable treasures” central to Japan’s global identity.
Under the new AI Promotion Act, which took effect in September 2025, Japan has laid down ethical rules for AI use in creative fields. Although the law currently lacks strict penalties, it allows the government to demand technical transparency from AI developers if violations persist. Officials have already hinted at invoking Article 16, which could compel OpenAI to disclose details about Sora 2’s filtering and moderation systems.
Along with that, the Japanese government has also urged OpenAI to move from an “opt-out” system, where copyright holders must manually exclude their work, to an “opt-in” approach that requires explicit consent before using protected content for training or generation. Authorities said creators deserve both control and compensation when their intellectual property is involved.
In response to the issue, OpenAI has stated that rights holders can opt out of Sora 2’s training data and outputs, but hasn’t clarified how this process works for Japanese studios. Though the lack of transparency has drawn criticism from local creators and gaming giants like Nintendo, who are concerned about their characters being portrayed in unauthorized or inappropriate settings.
