TOKYO, Japan – The Tokyo District Court has ruled for the first time in Japan that a person’s voice can fall under the protection of publicity rights, establishing a legal framework for addressing unauthorized commercial use of recognizable voices as generative artificial intelligence makes voice imitation increasingly accessible.
The ruling came in a lawsuit brought by Japanese voice actor Kenjiro Tsuda, who sought the removal of TikTok videos that he said used generative AI to imitate his distinctive voice. The court said a human voice can be regarded, like a person’s likeness, as a symbol of individual personality and that unauthorized use can infringe publicity rights when the voice’s commercial appeal is being exploited.
The court nevertheless dismissed Tsuda’s request for the videos to be removed because the account and the relevant video data had already been deleted. Crucially, the judgment did not make a final determination that the specific videos had infringed Tsuda’s publicity rights.
Court establishes protection for voices
Presiding Judge Aya Takahashi wrote that a human voice is, like a portrait, a symbol of an individual’s personality.
The court further stated that unauthorized use of a performer’s voice may constitute a publicity-rights infringement when the voice has the ability to attract customers and the use is primarily intended to exploit that attraction.
That distinction is central to the ruling. The judgment does not create a blanket prohibition on artificial voices or every instance in which an AI-generated voice resembles a real person.
Instead, it provides a legal standard for circumstances in which a recognizable voice has commercial drawing power and that value is being used without authorization.
Case involved 188 TikTok videos
Tsuda, known for his work as a voice actor and for voicing Kento Nanami in Jujutsu Kaisen, filed the lawsuit after an anonymous TikTok account posted videos using narration that he said had been generated to imitate his voice.
■ Related News
According to the case record and reporting on the judgment, 188 videos were posted between July 2024 and September 2025. The videos covered subjects including urban legends and paranormal material, while the account had accumulated roughly 210,000 followers. Viewers posted comments noting similarities between the narrator’s voice and Tsuda’s.
Tsuda’s side argued that the distinctive qualities of his voice had been intentionally reproduced with generative AI and that the use exploited the voice’s ability to attract viewers.
TikTok’s position differed. The company argued that the narration used a generic male voice and that the person who posted the material had described the AI system as having been trained using a friend’s voice.
Court did not decide whether Tsuda’s rights were infringed
The outcome is therefore more legally specific than a simple finding that Tsuda “won” an AI voice-cloning case.
The court determined that voices can fall within publicity-rights protection and set out circumstances in which unauthorized use could constitute infringement. But because the disputed videos had already disappeared, the court did not need to determine whether the particular material actually satisfied those infringement requirements.
The judgment records that the account was deleted and that the relevant data had subsequently been removed from the operator’s servers. The court therefore rejected the request for deletion because there was no remaining material for the defendant to remove.
This means the ruling establishes an important legal interpretation without awarding Tsuda damages or ordering TikTok to remove the videos.
Japan had already been examining AI voice protections
The decision follows a separate Japanese government examination of how existing civil-law protections should apply to unauthorized uses of people’s images and voices generated with AI.
Japan’s Justice Ministry established an expert panel in April 2026 to examine civil liability involving unauthorized use of portraits, voices and related personal characteristics. The ministry published its report on Aug. 7, saying the review addressed publicity rights, personality-related protections, damages, injunctions and the application of the Unfair Competition Prevention Act to generative-AI cases.
The ministry’s work specifically considered the growing problem of unauthorized AI-generated reproductions of voices and images. Its published materials examined circumstances in which generated content could implicate publicity rights and other protections.
The timing gives the Tokyo judgment particular significance: the court has now provided a concrete judicial interpretation in an actual dispute involving an AI-generated imitation of a performer’s voice.
Ruling does not amount to an AI voice-cloning ban
The decision should not be interpreted as a Japanese ban on AI voice cloning.
The court’s reasoning focuses on the circumstances in which a person’s voice possesses commercial attraction and an unauthorized use is primarily directed at exploiting that value. It does not establish that every synthetic voice resembling a real person violates Japanese law.
Japan’s Justice Ministry’s earlier examination similarly distinguished between potentially unlawful uses and forms of expression such as entertainment or imitation that may fall outside publicity-rights infringement depending on the circumstances.
The legal question will therefore remain fact-specific, including issues such as how recognizable a voice is, whether it has commercial drawing power, how the AI-generated voice is used and whether the purpose of the use is to capitalize on that identity.
Significance for performers and AI developers
The ruling gives Japanese performers a judicially recognized route for asserting that the commercial value of their voices can receive legal protection.
That is particularly relevant to voice actors and singers whose professional identities depend heavily on distinctive vocal characteristics. It also has implications for companies developing or deploying synthetic-voice systems, advertisers, entertainment producers and social-media platforms that host AI-generated content.
At the same time, the decision leaves important questions open because the court did not determine whether the disputed videos themselves constituted an infringement.
The immediate precedent is therefore narrower than a general “voice right” statute: Japan’s existing publicity-rights framework can encompass a person’s voice when the legal conditions identified by the court are met.
A precedent shaped by the AI era
The Tokyo judgment is the first reported Japanese judicial decision to recognize protection for a person’s voice under publicity rights in a case involving generative AI.
Its significance lies less in the deletion dispute itself than in the legal principle articulated by the court. As generative AI makes convincing voice imitation easier to produce and distribute, the judgment provides Japanese courts, performers and technology companies with a clearer framework for considering when a recognizable voice has legally protected commercial value.
The case does not resolve every question surrounding AI-generated voices. But it establishes that, under Japanese publicity-rights law, a human voice can have legally protectable commercial value separate from the person’s physical likeness.
Reporting Credit: Tokyo District Court and official judgment; Japan Ministry of Justice; Kenjiro Tsuda’s legal representatives as reflected in court reporting; TikTok/ByteDance statements and court submissions where reported. Secondary reporting from Associated Press, Jiji Press/Nippon.com, Reuters and The Guardian was used for chronology and case background.
