Masako Nozawa has been Goku for over forty years. She voiced the character since Dragon Ball’s 1986 debut, along with Gohan and Goten, building one of the most recognisable performances in anime history. Now 89, she is watching AI tools reproduce that lifetime of work without her consent — and she is not staying silent.
In comments reported by NHK on October 15, Nozawa watched an AI-generated clip imitating her voice and condemned the practice outright. “This is not my voice,” she said. “Using it without permission is unforgivable. It’s the same as theft, and it makes me angry.” Her objection landed the same day the Japan Actors Union — where she serves as vice-chairperson — published its first formal guidelines on AI voice use.
What the Guidelines Require
The Japan Actors Union represents roughly 2,600 actors and voice performers. Its new framework treats consent as non-negotiable at every stage of the AI production chain. Performers must give explicit written permission at three distinct points: voice use for AI training or analysis, voice model generation, and publication or sale of the resulting audio.
The guidelines go beyond a simple opt-in. They call for separate compensation at each stage, including specific payments for advertising applications and multilingual use. Buyout contracts are prohibited, and consent generally lasts one year. The union also restricts the use of deceased performers’ voices in new works unless heirs or authorised rights holders grant permission.
Not law: The guidelines are voluntary industry standards, not legislation. Japan does not yet have comprehensive statutory protections covering voice use in AI, though a Tokyo court ruling in September found that voices can fall under publicity rights.
On dubbing, the union states that AI should not replace human performers in film or television — a direct response to concerns that synthetic voices could undercut working actors across the industry.
‘It Lacks Heart’
Nozawa’s objection is as much about craft as it is about consent. She described the AI-generated version of her voice as technically impressive but emotionally hollow. “An AI-generated voice might sound like me, and it might even be an exact replica,” she said. “But it lacks emotion. Without emotion, it’s useless, no matter what the role is.”
She is not hostile to the technology itself. In a video message supporting the guidelines, Nozawa described AI’s progress as exciting and said she hopes humans and AI can “have fun together” — provided everyone follows the rules. Her position is specific: unauthorised use, not AI itself, is the problem.
“My voice is mine, as are my performances. If that’s going to be used, even a little bit, they should ask for permission.” — Masako Nozawa
The distinction matters. Nozawa has voiced Goku since she was roughly 50 years old. That performance history functions less like a job and more like a cultural inheritance — cloning that voice resembles training a model on a master chef’s unpublished recipes and then selling the output under their name.
A Court Ruling With a Catch
The legal landscape is shifting, if slowly. On September 30, the Tokyo District Court issued what is believed to be Japan’s first ruling on AI voice cloning, in a case voice actor Kenjiro Tsuda brought against TikTok’s operator over videos that used an AI copy of his voice.
The court found that a voice can fall under publicity rights where the use exists solely to exploit its commercial appeal, with presiding judge Aya Takahashi likening a voice to a likeness as a symbol of personality. But the court dismissed Tsuda’s takedown claim because the poster had already deleted the videos — leaving nothing for TikTok to remove.
The Enforcement Gap
Important questions remain unresolved. The union’s guidelines depend on companies choosing to follow them. Cross-border enforcement is murky — AI voice tools require no studio, no contract, and no credits, and can be operated from anywhere. Platform responsibility is equally unclear.
In August 2026, voice actress Megumi Ogata, known for portraying Shinji in Evangelion, described AI voice cloning as “heartbreaking” and called on the Japanese government to establish rules. Her comments followed a Ministry of Justice draft proposing protections for voices under publicity rights.
Nozawa said she hopes the new guidelines mark the beginning of enforceable standards. “Rules are now being established,” she said. “I hope everyone follows them.” For the thousands of voice actors whose livelihoods depend on the distinctiveness of their performances, the question is no longer whether AI can replicate a voice — it’s whether anyone will be required to ask first.
