Tokyo Court Says Voices Get Publicity Rights in Kenjiro Tsuda's AI Clone Case

TL;DR

On September 30, the Tokyo District Court held that a human voice, like a person's likeness, can be protected under publicity rights. That is the first time a Japanese court has said so. It came in a suit by voice actor Kenjiro Tsuda against TikTok's operator over an anonymous account that posted at least 188 videos narrated by an AI voice that sounded like him. Tsuda's actual request, a takedown, was dismissed because the account and its data were already gone. He lost the case and won the principle, which for everyone selling or using synthetic voices is the part that matters.

What happened

Tsuda is one of the most recognizable voices in anime: Kento Nanami in Jujutsu Kaisen and Seto Kaiba in Yu-Gi-Oh!. His low, husky delivery is a brand in its own right, which is exactly why someone wanted it.

According to Nikkei and ITmedia, an anonymous TikTok account posted at least 188 videos between July 2024 and September 2025. The subjects were urban legends, trivia and occult stories, each with a generated narration that Tsuda's side said copied his voice. The account built more than 210,000 followers and, by the estimate cited in the suit, earned the uploader 500,000 to 750,000 yen a month.

Tsuda sued TikTok's operator in November 2025 under two theories: publicity rights and Japan's Unfair Competition Prevention Act. He asked for one thing, deletion of the videos. He did not seek damages, Music Business Worldwide reports.

What the court actually held

The core finding: a person's voice, like their likeness, is a symbol of their personality, so it can carry the same publicity-rights protection Japanese courts already give to faces and names. Presiding judge Takahashi put it this way, as quoted by MBW: "The unauthorized use of a performer's voice ... can be considered an infringement on publicity rights."

The protection is not a blanket ban on sounding like someone. As coki reads the judgment, infringement needs a voice that carries real customer-attracting power as a performer's, and a use whose main point is to exploit that pull for commercial gain. That framing tracks the Supreme Court's 2012 Pink Lady decision, which set the test for when using a celebrity's likeness crosses the line.

Why the takedown still lost

The account holder deleted the account on May 29, 2026. Nobody asked to restore it within the 30-day window, so the data was wiped from TikTok's servers by the end of June, ITmedia reports. A court cannot order a platform to delete what it no longer holds, so the claim failed on mootness.

That also means the court never decided the hard question: whether these 188 videos actually used Tsuda's voice. TikTok argued the narration was a "generic male voice" and pointed to the uploader's claim that the model was trained on a friend's voice. Coki notes the judgment did not weigh the acoustic analysis Tsuda's team submitted, and the unfair-competition theory went unaddressed too.

Think of it as a court confirming your house is protected by trespass law, then declining to rule on the burglar because he already moved out and took the evidence with him.

Why builders should care

Japan just got a voice right without a new statute

Japanese law never spelled out a right in one's voice, and there was almost no case law on it. This ruling builds it from existing doctrine instead. That makes the theory available to every Japanese performer right now, not after a legislative cycle. Waseda professor Tatsuhiro Ueno told Nikkei it supports protection of voice actors' rights, while warning that proving an AI voice matches a specific person is harder than proving a photo shows someone.

The risk sits with the monetized sound-alike

If you ship a TTS product, a voice library, or a content pipeline that targets Japanese audiences, the exposure is clear: a voice that is recognizably a famous performer's, used to pull views or sell something, is now a publicity-rights problem in Japan. "We trained it on a friend" is a factual defense you would have to win at trial, not a magic phrase. The account here was plainly using the voice for reach, down to a profile picture that resembled a Jujutsu Kaisen character, according to MBW.

Platforms got a pass on timing, not on principle

TikTok walked out without an order against it, but only because the content had vanished first. A future plaintiff who sues while the videos are still up has a ready-made legal theory and a court that has already accepted the premise. Expect faster filings and earlier preservation requests.

The economics explain the incentive

The reported 500,000 to 750,000 yen a month is roughly $3,200 to $4,800. Over about 14 months of uploads, that is a back-of-envelope 7 to 10.5 million yen for an anonymous account whose main asset was someone else's throat. Nobody needed a frontier model to do this, which is why it will keep happening until it gets expensive.

What is still open

• Similarity proof. No court has yet ruled how close an AI voice must be to count as someone's voice.
• Platform liability. The ruling did not decide what TikTok owed while the videos were live.
• Damages. Tsuda asked for none, so there is no number yet for what a cloned voice is worth in court.
• Unfair competition. The second legal theory was left untouched.

The Japan Actors Union said an actor's voice reflects years of training and deserves basic legal protection, per Digital Music News. Japan's voice-acting industry has been pushing on unauthorized AI generation for a while, and this is the first time a court has put that argument on paper, as Automaton notes.

Key Takeaways

• The Tokyo District Court ruled on September 30, 2026 that a voice can be protected by publicity rights, a first for Japan.
• Infringement requires a voice with real commercial pull and a use aimed mainly at exploiting it, mirroring the 2012 Pink Lady likeness test.
• Tsuda's takedown claim against TikTok's operator was dismissed because the account was deleted in May and its data purged in June.
• The court did not decide whether the 188 AI-narrated videos actually copied Tsuda's voice, or whether TikTok was liable.
• Anyone monetizing famous-sounding synthetic voices for Japanese audiences now faces a recognized legal theory, without waiting for new legislation.

Sources: Music Business Worldwide, ITmedia NEWS, Nikkei, coki, Digital Music News, Automaton, Monolith Law (Pink Lady case)