On October 18, the Japanese Society for Rights of Authors, Composers and Publishers (JASRAC) attended and expressed its opinions at the public hearing of interested parties held during the second session of the Cabinet Office's Committee to Discuss Intellectual Property Rights in the Age of AI. The following is a summary of its opinions.


1. Relationship between generative AI and copyrights
JASRAC believes that the development and use of generative AI can benefit both creators and cultural development if they are in harmony with the creative cycle.

However, in relation to copyrights, JASRAC cannot agree to separating the learning aspect from the generation/use aspect of generative AI, and determining that there is no general issue with the former because no enjoyment of expressions of works takes place, and that there are no issues regarding the latter as long as proof of dependence/similarity are grounds for copyright infringement.

The International Confederation of Societies of Authors and Composers (CISAC), which comprises 225 collective management organizations in the field of copyright from around the world, has pointed out the following problems regarding TDM (Text and Data Mining) learning:

Music creators in Japan are voicing similar concerns. Especially with regard to point (2) above, they express their concerns and sense of crisis that it would not be fair that the competition between generative AI and creators would lead to negative consequences, as opposed to the competition among creators.


While the Copyright Act alone will not remove the concerns of music creators and other right holders, JASRAC is asking the government to consider the following four points in order to alleviate their concerns:

2. Technical measures to avoid risks to intellectual property rights concerning generative AI
JASRAC considers it preferable that right holders' concerns and disadvantages be resolved through technological measures. In particular, technologies that would help ensure transparency and identify AI-generated works would be useful. However, as it is also foreseeable that technologies to neutralize such solutions will also be developed, technological solutions will have their limits.

Technologies that enable creators to choose not to have their works subject to AI learning (i.e., opting out)are limited as well in that, for musical works, content data is often distributed by third parties other than creators. In the absence of legal guarantees to the contrary, utilization of that technology in a way that respects the creators' intentions cannot be expected.

Thus, JASRAC considers it necessary for the government to consider legal measures (such as those listed in 1. (i) through (iv) above) as well as technological measures.


3. Remunerating creators in relation to generative AI
While the Committee to Discuss Intellectual Property Rights in the Age of AI illustrates how profits may be returned to creators through the provision of learning data for a fee, JASRAC considers that it would be difficult for music creators, such as authors and composers, to individually enter into such contracts with generative AI developers. Furthermore, since creators working at the front line of the music industry themselves do not necessarily own rights to sound recordings (i.e., the rights of producers of phonograms), it would also be difficult for them to provide their music data to generative AI developers for a fee by registering the data with a stock content service.

Given the existence of the copyright exception, it is difficult to simulate practical means of remuneration to music creators. Therefore, it is necessary to consider legal measures (such as those listed in 1. (i) through (iv) above) as well as practical means of remuneration to music creators.

To achieve the continuous development of culture and art and associated businesses, JASRAC believes that it is critical to ensure an environment where creators can concentrate on creative activities without anxiety or concern. JASRAC will continue its own research and make suggestions with the aim of achieving a framework for AI utilization that is in harmony with the creative cycle.