Japan's Copyright Challenge to Sora 2 - Japanese Creative Giants Demand Halt on Sora 2 Training
The international dispute over copyright infringement in generative AI has intensified. That is due to a direct and unprecedented legal challenge issued by major Japanese content creators against OpenAI concerning its advanced video-generation model, Sora 2.
The Content Overseas Distribution Association (CODA), which represents leading Japanese intellectual property owners including Studio Ghibli and Bandai Namco, has formally requested that OpenAI immediately cease using their members' content for the training and operation of Sora 2.
That is a request founded on direct observations. CODA states they have confirmed that a significant portion of content generated by Sora 2 closely resembles Japanese copyrighted images and stylistic elements. That is strong evidence suggesting the unauthorized use of Japanese intellectual property as machine learning data.
The Legal Conflict: Opt-In vs. Opt-Out
The core of the dispute highlights a fundamental clash between OpenAI's policy and Japanese law. While Sam Altman (CEO of OpenAI) stated that OpenAI would allow content owners to opt-out of having their work used for Sora 2 training, CODA asserts that this policy is illegal under Japanese jurisdiction from the outset. They are emphasizing that under Japanese copyright law, prior permission from the copyright holder is generally required for the use of intellectual property. That is no system allowing one to avoid liability for infringement through subsequent actions.
The Japanese government previously issued a similar appeal. However, CODA's letter represents a direct, unified action from the creative sector, demanding that OpenAI "respond sincerely" and respect the nation's copyright legislation. That is a crucial legal battle expected to define the global standards for AI training data ethics.
Reference:
