TOKYO - Efforts to harness the voices of dubbing artists for commercial use are gaining momentum in Japan, spotlighting their potential to boost the global reach of Japanese anime in the age of artificial intelligence.
AI-powered tools are also expected to strengthen protections for performers whose voices have become increasingly valuable even as their professional status and rights remain vulnerable.
The Tokyo-based Voice Integrity and Dubbing Advancement Association unveiled its initiatives toward these goals at a press conference last November.
Voice actor Mika Kanai took the stage and delivered a simple line in Japanese: "You just said you are hungry, but what exactly do you want to eat?" Her words were then rendered in multiple foreign languages, including English and Chinese -- but they still sounded like Kanai.
The demonstration showcased a generative AI system developed by U.S. software firm ElevenLabs, which specializes in AI voice generation. The technology can translate spoken words into more than 30 languages while preserving the speaker's original voice quality, according to VIDA.
The system also embeds digital watermarks into media files, allowing creators and rights holders to track how voice recordings are used, verify ownership and help prevent unauthorized copying or theft.
The association, founded by voice actor agencies and other industry players, plans to use the technology in consultation with rights holders to support the overseas expansion of Japanese anime.
"Sales of Japanese anime have increased significantly overseas, and voice actors are very popular," said VIDA representative director Masakazu Kubo. "The number of fans will further increase if voice actors' original voices are used without dubbing."
In a related development, major trading house Itochu Corp. and the Japan Actors Union announced an agreement last November to establish an official voice database known as J-Vox-Pro.
The platform is intended to safely store, manage and provide performers' voices -- including those of voice actors and other performers -- for use in Japan and overseas.
The initiative reflects rising concern within the industry about how easily voices can now be reproduced using AI.
Under current Japanese law, a person's voice itself is not recognized as copyrighted material. As a result, AI systems can legally learn from recordings of voice actors and generate similar-sounding voices without violating copyright rules.
Videos using such AI-generated voices without permission are already widely circulated online.
For many performers, the issue carries economic implications as well as legal ones.
Voice actors often struggle to find steady work, with many earning less than 3 million yen (about $19,000) a year.
Because they often work as freelancers, their personal and economic rights can be difficult to protect, according to Itochu and other organizations involved in the project.
The J-Vox-Pro database is designed to create a mechanism through which performers can earn income when their registered voices -- or AI-generated voices based on them -- are used by companies.
Potential applications for such technology extend beyond entertainment. AI-generated versions of dubbing artists' voices could be used in car navigation systems or to read medical instructions for elderly people who have difficulty seeing small print.
The Japan Actors Union is also considering additional measures to combat the unauthorized use of performers' voices, including seeking court injunctions against unapproved reproductions and suing for damages.
"The private-sector initiative to certify the original voices of dubbing actors should be effective in preventing the use of unauthorized voices to some extent," said Kazuhiro Ando, a professor of intellectual property law at Toyo University.
"But as there are limits, it is necessary to advance legislative measures and crack down on the practice," he said.
