Key Points
- The Japan Actors Union has unveiled new guidelines restricting generative AI from replacing voice actors in dubbing and barring the use of deceased performers' voices without family consent.
- The initiative follows rampant voice cloning on social media and a landmark Tokyo court decision recognizing a celebrity's voice as legally protected intellectual property under publicity rights.
- Under the rules, creators must secure explicit consent from performers across every step of AI voice modeling, synthesis, and commercial deployment.
Summary In response to a rapid surge in unauthorized synthetic voice generation, Japanese performing arts groups have drawn a firm line against the misuse of generative artificial intelligence in the entertainment industry. The Japan Actors Union, which represents prominent screen actors and voice talents (seiyuu), held a press conference in Tokyo to announce comprehensive guidelines aimed at safeguarding performers' vocal rights against unregulated AI replication.
Under the new guidelines on generative AI voice usage, the exploitation of a deceased performer's voice for newly produced works is strictly prohibited in principle, permitting exceptions only with explicit authorization from bereaved families or legal estates. Furthermore, to safeguard performers' livelihood and professional employment opportunities, the union declared that generative AI should not be deployed to replace voice actors in dubbing feature films and television dramas.
The union's protocol establishes mandatory consent at every stage of voice generation: initial data collection and analysis ('learning and interpreting'), the synthesis phase that reconstructs tone and vocal timbre, and the commercial distribution or sale of synthesized audio. Performers retain the right to re-evaluate consent once initial contracts expire. At the press briefing, veteran voice actor Bin Shimada, widely recognized for his roles in iconic animation series like 'Chibi Maruko-chan', voiced his personal anguish, recalling his shock and mounting anger upon discovering unauthorized AI audio circulating online that cloned his voice singing songs as an anime character.
Unauthorized voice cloning and unauthorized likeness use have escalated dramatically across East Asian media and social networks. According to a 2025 survey by a non-profit watchdog monitoring publicity and likeness rights, unauthorized celebrity likenesses exceeded 40,000 posts on social media platforms, racking up more than 335 million cumulative views. The union's guidelines arrive on the heels of a landmark ruling by the Tokyo District Court in favor of renowned voice actor Kenjiro Tsuda. Tsuda had initiated legal action against a TikTok account using an unauthorized AI reproduction of his signature deep, baritone voice to narrate horror stories and conspiracy theories, leading the court to establish that a celebrity's distinct voice constitutes protected property under legal publicity rights.
Sponsored · Ad