Japanese Government Panel to Focus on Generative AI, Copyright Concerns in Discussions

Reuters file photo
The logos of Google, Microsoft and Alphabet are seen in this illustration.

A government subcommittee has begun discussions on copyright concerns and generative artificial intelligence, which can create text and images using data from the internet.

Under Article 30-4 of the Copyright Law, copyrighted materials can be used in the development and training of AI without the permission of copyright holders, unless “the action would unreasonably prejudice the interests of the copyright owner.”

The subcommittee of the Cultural Affairs Agency’s Culture Council, which started substantial discussions at its meeting Wednesday, said it will discuss the criteria for such exceptions, among other issues.

Creators such as illustrators are concerned that their work is being used without their permission to train AI models.

Policies governing the use of copyrighted materials to train AI models will be a key part of the subcommittee’s discussions. The subcommittee will also discuss what kinds of AI-generated content constitute copyright infringement and what kinds can be recognized as copyrighted materials.

The Cultural Affairs Agency said, while the subcommittee is considering discussion points, the agency will also exchange opinions with creators and AI firms.

On Wednesday, some subcommittee members said concerns raised by copyright holders and creators should be the focus of discussions. Members also noted that technical factors would be an important element.