Japan recognizes a voice as a publicity right for the first time, then dismisses the claim it was built for
On 30 September 2026 the Tokyo District Court became the first Japanese court to hold that a voice can be protected under the right of publicity, in a suit voice actor Kenjiro Tsuda (Kento Nanami in Jujutsu Kaisen) brought against TikTok over an anonymous account that posted at least 188 AI-narrated videos between July 2024 and September 2025 in a voice viewers said resembled his. The court dismissed his claim anyway: by the time of judgment the videos were already gone, so it never decided whether they actually infringed.
A voice now gets the same legal test as a face in Japan, and the first case under it produced a loss
Japan has no statute that names voice as a protected asset. The Tokyo District Court got there by extension: in its 2012 ruling in the case brought by the idol duo Pink Lady, the Supreme Court held that unauthorized commercial use of a person's name or likeness infringes publicity rights only when it is aimed exclusively at exploiting that person's power to attract customers, not whenever it is merely referenced. Presiding judge Aya Takahashi applied the identical test to voice on 30 September 2026, reasoning that a voice, like a portrait, is a symbol of individual personality. Reuters, AP and Jiji all flagged it as a first for a Japanese court.
TikTok's defense in the case was not that voice cannot be protected. It was that the narration was a "generic male voice" and that any resemblance to Tsuda was subjective, a defense that concedes the doctrine and fights the facts. The account operator had separately stated, on an external site, that the AI model behind the narration was trained on "a friend's voice," not Tsuda's, a claim the court never tested because it never reached the merits. Our read: the case establishes that the test exists, but supplies no finding at all on how close an AI clone has to sound, or how a court weighs a sworn claim about whose voice trained the model, before "exclusively commercial" is satisfied.
What the court actually decided, and what it pointedly left open
The suit named TikTok, the platform, as defendant, not the anonymous poster. Tsuda sought removal of the videos and likely damages tied to a separate claim, under Japan's Unfair Competition Prevention Act, that the poster earned an estimated JPY 500,000 to 750,000 a month by creating confusion with his voice. The judgment did not address that second claim at all. On the publicity-rights claim, the court's reasoning had two steps that matter independently: first, that a voice can be protected the same way a name or likeness can; second, that TikTok, specifically, had no further obligation to remove content that was no longer on its servers. The second step did all the work in the outcome. The first step is the one that makes this a first.
Tsuda's lawyer called the result a ruling that "ended up making both sides look victorious," which is an accurate description of a decision that grants a legal principle and denies every remedy that depended on it. TikTok told AFP it would "respect the court's judgment while considering what steps to take" after reviewing the details, language that commits it to nothing specific.
The numbers reporters disagree on, and the ones that are solid
| Quantity | Value | Source |
|---|---|---|
| Ruling date | 30 September 2026, Tokyo District Court | Orrick, AP, Jiji, Music Business Worldwide |
| Videos posted | "at least 188" (Orrick) / "more than 180" (AP) | Orrick client alert; AP via ny1.com |
| Posting period | July 2024 to September 2025 | Orrick |
| Account followers | "more than 210,000" (Orrick) / "more than 200,000" (AFP via multiple outlets) | Orrick; AFP wire |
| Alleged poster earnings | JPY 500,000 to 750,000 per month (Unfair Competition claim, not addressed by the ruling) | Orrick |
| Legal test applied | 2012 Supreme Court "Pink Lady" exclusivity test, extended from name/likeness to voice for the first time | Orrick; Music Business Worldwide |
| Outcome | Publicity right in voice recognized in principle; takedown claim dismissed because videos were already deleted; merits of infringement not decided | Orrick, AP, Jiji |
| Defendant's position | Narration was a "generic male voice"; resemblance "subjective" | Courthouse News, AP |
| Time from filing to ruling | approximately 10 months (filed ~November 2025) | Orrick; our calculation from reported filing month |
Values as reported in primary and wire coverage dated 1-6 October 2026; where two sources diverge, both figures are shown rather than one chosen silently.
The spread between "188" and "more than 180," and between 200,000 and 210,000 followers, is small, but it is a reminder that even a decision wire services agree is a legal first does not come with a single, matched set of case facts across reports. Neither figure changes the holding.
How the exclusivity test moved from a magazine cover to an AI narrator
The structural point for anyone building or deploying AI voice tools: a legal right can attach to a technology before any court has ever applied it to a contested fact pattern. Right now, Japan has a voice-publicity doctrine with zero merits rulings under it. Every subsequent case that resolves the same way, on mootness or deletion, extends that gap instead of closing it.
What this means for dubbing, voice cloning and localization operators in Japan
Three practical consequences follow directly from the two-step reasoning above. First, the exclusivity test's content is still undefined for anything short of outright impersonation-for-profit: a dubbed character voice, a synthetic brand narrator, or a cloned voice used with a disclosed AI label are all plausibly outside "exclusively commercial exploitation," but no ruling says so yet. Second, fast deletion is a working defense in Japan in a way it would not be in a jurisdiction that allows a claim for past harm to proceed regardless of present removal; an operator that pulls disputed content quickly, even after the fact, may avoid ever having a court examine whether it infringed. Third, an unverified claim about whose voice trained a model ("a friend's voice") was never tested here, which means Japanese case law currently offers no guidance on what evidence of consent, or its absence, would need to look like to win or lose a similar claim on the merits.
None of this gives an AI dubbing vendor a green light. It gives a specific, narrow fact pattern, no statute, a precedent extended by analogy, and a remedy question left open. Consent documentation for any voice used in training or cloning is the only thing in this record that would have mattered if the case had reached the merits, and it is exactly what the next case, if its content stays up long enough to be judged, will test.
Check it yourself
- Primary summary: Orrick client alert, "Tokyo Court Recognizes Publicity Rights for Voice," published 6 October 2026, by Hiroki Sugita (Tokyo). orrick.com/en/Insights/2026/10/Tokyo-Court-Recognizes-Publicity-Rights-for-Voice
- Wire coverage naming Tsuda and the Jujutsu Kaisen role: Music Business Worldwide, 2 October 2026, and The Next Web, both corroborating the AP/Jiji account.
- AP coverage: ny1.com, 1 October 2026.
- Jiji/nippon.com: nippon.com/en/news/yjj2026093000118, 30 September 2026.
- The 2012 Supreme Court precedent being extended is widely referenced as the "Pink Lady" ruling on unauthorized use of a celebrity's likeness in a weekly magazine; none of the sources above give its formal case citation, so we have not invented one.
What would prove this reading wrong
Our reading is that Japan now has a voice-publicity doctrine that has only ever been applied to a moot fact pattern, and that fast deletion functions as a practical shield against a merits ruling. That is falsifiable: if, by 31 December 2027, a Japanese court rules on the merits of a voice-publicity infringement claim where the disputed content was still live at the time of judgment, whichever way it comes out, our claim that this doctrine has produced zero merits rulings would be wrong, and this would become the case to read instead of Tsuda's.
Sources
- Hiroki Sugita, "Tokyo Court Recognizes Publicity Rights for Voice," Orrick, 6 October 2026. orrick.com
- Associated Press, "Tokyo court grants legal protection to human voices in AI clone case," via ny1.com, 1 October 2026.
- Jiji Press / nippon.com, "Japan Court Recognizes Publicity Rights in Voice for 1st Time," 30 September 2026. nippon.com
- Music Business Worldwide, "Tokyo court rules that a person's voice can be protected by publicity rights, a first for Japan," October 2026.
- Courthouse News Service, "Japan court rules human voice is protected in TikTok AI case," October 2026.