In a landmark decision, a Tokyo court has affirmed that a person’s voice is protected under the law, marking the first time such a ruling has been made in Japan’s judicial history. This pivotal judgment came during a lawsuit filed by renowned voice actor Kenjiro Tsuda against TikTok.
AI Voice Clones Infringe on Image Rights, Court Rules
The court partially sided with Tsuda, a veteran voice actor with over three decades of experience lending his voice to major anime series, films, and video games, including the iconic character Seto Kaiba in Yu-Gi-Oh!. Tsuda had sued TikTok, alleging that an anonymous account had used AI-generated cloned versions of his voice in multiple videos uploaded between July 2024 and late last year.
Tsuda argued that the AI-generated narration in the videos featured his distinctively deep and “sonorous” baritone voice, making it clearly identifiable as an imitation of him. His legal team highlighted that the account in question had amassed over 200,000 followers and attracted significant attention precisely because the novelty of a famous voice actor narrating “suspicious or vulgar content” drew viewers in.
The court’s ruling stated that using an actor’s voice without their explicit permission constitutes an infringement of their image rights. While the court acknowledged the validity of Tsuda’s claim regarding the misuse of his voice, it could not order TikTok to remove the videos as the offending account had already been deactivated.
TikTok’s Defense and Broader AI Concerns
TikTok, in its defense, had argued that the voice used in the videos was merely a “generic male voice” and that its similarity to Tsuda’s voice was debatable.
This case sheds light on the growing concerns within Japan’s vibrant creative industries, particularly among anime voice actors. Despite the Japanese government’s strong push for AI development at a national strategic level, creators are increasingly worried about the lack of adequate protections for their work and personal likeness, which are often used to train AI models.
Last year, the Japanese government formally requested OpenAI that content considered “irreplaceable treasures” of Japanese anime and manga should not be used in ways that infringe copyright for training generative models like Sora 2.
The ruling by the Tokyo court is a significant step towards providing legal recourse for individuals whose voices may be replicated and misused by AI technologies, underscoring the evolving legal landscape in the age of artificial intelligence.
