3 days ago
- Entertainment
- Yomiuri Shimbun
Unauthorized AI Generation: How Can Rights of Voice Actors, Singers Be Protected?
The act of using artificial intelligence to generate voices that sound identical to those of voice actors and singers and uploading them to the internet has been rampant. Discussions should be deepened on how to protect their 'voice rights.'
The internet is flooded with videos of voice actors speaking lines unrelated to works in which they participated and singers performing other singers' songs. It is believed that such content has been created by training generative AI without permission to create highly similar voices.
The Japan Actors Union, whose members include many voice actors, has reported that it confirmed about 270 such cases in an investigation that was conducted over a three-month period. The union and other organizations have called for obtaining consent from rights holders when AI is used to learn voices, and they have also urged people to clearly indicate when content is generated by AI.
For voice actors and singers, their voices are 'products' crafted through training and experience accumulated for years. If videos posted online get views and earn profits by using their voices, this act appears to constitute a clear infringement on their rights.
Works such as text and musical compositions are protected by the Copyright Law, but voices have been interpreted conventionally as not falling under this protection.
Celebrities and athletes are said to have the 'right of publicity,' which allows them to exclusively use their names and likenesses. However, whether this right also covers voices remains unclear.
The Economy, Trade and Industry Ministry has warned that acts — such as a case in which voices of voice actors and other related parties were created by AI without their consent to produce alarm clocks and such products were sold — could constitute a violation of the Unfair Competition Prevention Law.
However, some people argue that applying this law is difficult because proving a violation requires the AI-generated voice to be immediately recognizable as the voice of a specific person, among other requirements.
In the U.S. state of Tennessee, a law was enacted last year to protect individuals' rights to their voices and likenesses from digital reproduction. It is hoped that Japan will explore ways to institute legal protection by referring to such instances.
In the process of considering that, various issues will likely arise, such as how to handle impersonations by entertainers. Certain considerations may be necessary to prevent a decline in forms of cultural expression such as parody.
Japan amended the Copyright Law in 2018 to allow AI tools to learn text and music without permission. As a result, the act of having AI learn animation and other works has become legal, creating a situation in which voices have been used freely without consent from rights holders.
It is necessary to amend the law again to stave off unauthorized AI training in the first place.
The use of AI has also raised concerns over issues such as fake videos featuring politicians that could influence elections and the spread of sexually explicit fake images, including child pornography. Measures against these deepfakes, among other problems, also must be implemented urgently.
(From The Yomiuri Shimbun, July 27, 2025)
