Japan’s animation studio Studio Ghibli has asked OpenAI, the developer of ChatGPT, to stop unauthorized artificial intelligence (AI) training on its content.

According to the U.S. information technology (IT) outlet TechCrunch on the 3rd (local time), Japan’s Content Overseas Distribution Association (CODA), which counts Ghibli among its members, sent an open letter to OpenAI, which recently launched the video-generation service ‘Sora 2’, asking it not to use member companies’ content for training without permission.

CODA also demanded that it respond in good faith to members’ complaints and inquiries regarding Sora 2’s copyright infringement. CODA pointed out, “We have confirmed that Sora 2 generates a large volume of videos similar to existing Japanese content,” and, “In situations where, as with Sora 2, outputs are reproduced or generated in a way similar to particular works, acts of reproduction during the training process may constitute copyright infringement.”

CODA went on to say, regarding OpenAI’s adoption of an ‘opt-out’ approach that has rightsholders request a ban on use of their content, “Under Japan’s copyright law, prior permission is required when using a work,” and, “There is no system that allows one to avoid liability for infringement through after-the-fact objections.”

Since OpenAI released Sora 2 at the end of September, users have been sharing videos that use AI to generate popular brands and animated characters. Around March, news that ChatGPT could generate images in Ghibli’s art style also spread, sparking a global trend of users converting their own photos. CNBC, the U.S. business news channel, predicted that Sora 2 would face a series of copyright lawsuits.

However, TechCrunch reported that, in the United States, there are currently no precedents interpreting how judges will apply copyright law to AI training and reproduction.