Japan Actors Union Develops Guidelines for Generative AI Use
The Japan Actors Union (JAU) presented the first version of its new guidelines for the use of generative AI during a press conference on Thursday. The 21-page "Guidelines for the Use of Generative AI on Vocal Audio Work" ("Onsei ni Kansuru Saisei AI Riyō Guideline”) is available as a PDF file on the JAU website with six sections and 22 articles:
- Section 1: General Provisions:
- Article 1: Purpose
- Article 2: Defining Terms in These Guidelines
- Article 3: Criteria for Unauthorized Generative AI Use in Vocal Audio
- Article 4: Guidelines' Scope of Application
- Section 2: Standards for AI Use
- Article 5: Standards for Prioritizing Human Performers
- Article 6: Concerning the Use of Vocal Audio Data of Deceased Persons
- Section 3: Rules for AI Usage
- Article 7: Prior Notification and Information for the Use of AI
- Article 8: Requirements for Clear, Written Consent for the Use of AI
- Article 9: Requirements for Notifications and Disclosures by Businesses
- Article 10: Rights to Request Crediting and Removal
- Article 11: Prohibitions on Unfavorable Treatment for Refusing or Withdrawing Consent
- Article 12: Period of Validity for Consent
- Article 13: Restrictions on Use and Prohibited Acts
- Section 4: Special Provisions
- Article 14: Special Provisions for Japanese-Language Vocal Audio for Animation and Dubbed Content
- Article 15: Special Provisions for Multilingual Localization and Overseas Distribution of Vocal Audio Works
- Section 5: Compensation, Oversight, and Other Measures
- Article 16: Compensation for Training and Usage
- Article 17: Data Oversight
- Article 18: Auditing and Data Log Requesting Rights
- Article 19: Requesting Cease and Desist, Removal of Unauthorized Generative AI Vocal Audio Works
- Article 20: Legal Action Towards Violations
- Section 6: Miscellaneous Provisions
- Article 21: Succession of Rights and Obligations
- Article 22: Regarding Revisions to the Guidelines
The guidelines define vocal audio as all audio recordings of actors and voice actors, including private audio recordings. They further note that AI use includes the use of various AI technologies, voice synthesis technology, deep learning, machine learning, and modeling via prompts, tagging, configuring metadata, operating GUI screens, and file specifications.
In JAU's press conference for the generative AI guidelines, Dragon Ball voice actors Masako Nozawa (Son Goku) and Bin Shimada (Broly) gave their thoughts on the issues. Nozawa said, “I'm excited the world we dreamed of when I was a child is slowly becoming a reality. But our voices are being used without our consent. It's a bit disappointing.” The voice actor continued, “I hope by everyone following the rules, we can create a world where humans and AI can have fun.”
In turn, Shimada noted he had heard a generative AI version of Ado's song “Usseewa” using his voice of Broly, as heard in the Dragon Ball Z franchise. “I was stunned the first time I heard it. I let out a deep sigh. And after some time, I grew angrier.” He added, “Enjoying it for personal use is OK. But, no matter what, I can't accept a voice actor's voice being released online without their knowledge, saying or singing something they haven't, to gain followers and profit.”
Japanese voice actors have been dealing with unauthorized use of their voices for several years. In his own bid to protect his voice from unauthorized generative AI use, voice actor Yūki Kaji developed his own voice synthesis program Soyogi Fractal in June 2024. Prior to Soyogi Fractal's release, Kaji stated he was concerned about copyright issues surrounding voices, which led to the development of the voice synthesis program. Kaji then left the talent agency VIMS on April 8 and started his own company for Soyogi Fractal the next day.
In late October 2024, several voice actors announced they had founded the No More Mudan Seisei AI (No More Unauthorized Generative AI) group. The organization was founded to protect voice actors from unauthorized generative AI use of their voices.
Voice actor Kenjirō Tsuda filed a lawsuit last November against the operator of the video-sharing app TikTok for posting videos that allegedly used AI to imitate his voice without permission. In September, the Tokyo District Court dismissed the lawsuit after the videos had been removed. Both an expert panel for the Ministry of Justice and the court affirmed that the voices of voice actors and actors are legally protected under the right of publicity, just as their likenesses (such as photographs) are.
Sources: Japan Actors Union's website (link 2), PR Times via Eiga Natalie
KioskNews shows a cleaned-up reading view extracted from the publisher’s page — the original always lives on their site, not ours.