Tokyo Court Protects Voices—but Leaves Tsuda’s AI-Cloning Claim Undecided

The September 30, 2026 decision matters for what it recognizes, not for a penalty it imposed. It brings a commercially valuable part of a performer’s identity within legal protection. It is neither a court-ordered takedown in this case nor a finding that every unauthorized synthetic imitation is unlawful.⁠17

Presiding Judge Aya Takahashi treated the human voice as an expression of personal identity comparable to a likeness. Tokyo Shimbun reproduced her reasoning in Japanese; the following is a translation of that reported passage, not wording independently checked against the complete judgment.⁠4

The Guardian and Japanese legal-news reporting describe the decision as Japan’s first express judicial recognition that voices are protected by publicity rights. That is a significant legal development, but recognition of a protected interest and proof of infringement are two different questions. This case answered the first and left the second undecided.⁠1⁠6

The condition is specific: unauthorized use of a performer’s voice can infringe publicity rights when it is exclusively directed toward exploiting the voice’s commercial appeal—its power to attract customers and promote sales. The word exclusively is not incidental. A recognizable voice and a lack of permission do not, by themselves, establish that this test has been met.⁠17

That leaves important boundaries unresolved. The available account does not settle how the court would treat parody, commentary or other particular uses. Nor is this a blanket ban on synthetic voices or artificial intelligence. Without the full judgment, it would be premature to turn a reported legal principle into a comprehensive rule for every kind of imitation.

Tsuda’s case concerned 188 videos posted to an anonymous TikTok account between July 2024 and September 2025, according to September 30 ORICON reporting. He alleged that they used an AI-generated copy of a voice that could only be his. That is his identification claim, not a technical finding about how the audio was produced. No verified account identifies the uploader, the AI tool, its training material or the method of imitation.⁠1⁠14

The voice has professional value as well as recognizability. Tsuda plays Kento Nanami in Jujutsu Kaisen and Hyakunosuke Ogata in Golden Kamuy; his credits also include Yu-Gi-Oh! AFP describes his “lustrous” baritone. For a performer whose sound is part of the work audiences seek out, an imitation can trade on something more consequential than familiarity.⁠16⁠20⁠24

Tsuda’s legal team characterized the account’s content as “sordid,” according to The Guardian. That description belongs to the lawyers: it is not a judicial finding of criminal obscenity. Likewise, the alleged imitation should not be treated as technically proven simply because the court recognized that voices can receive protection.⁠1

The account operator deleted the videos after the lawsuit was filed, and the removal included the account itself. With the disputed material gone, the court dismissed the demand against TikTok’s operating company to delete it. Dismissal therefore does not mean the court rejected protection for voices.⁠6

Tsuda’s lawyer, identified in Bengoshi.com reporting as 平野敬, said, in translation: “All the videos were deleted along with the account, so our original objective was fully achieved.” The available reporting does not establish an English spelling of his name or an appeal plan. It also establishes no damages demand or award, or separate prospective injunction.⁠6

This distinction is obscured by The Guardian’s September 30 headline, which says the copied voice “violated his rights.” ITmedia’s account published that day explicitly says the court did not decide whether these videos constituted infringement. The detailed Japanese reporting supports the narrower outcome: voice protection was recognized, but liability for the removed videos was not adjudicated.⁠1⁠17

The reported scale explains why the commercial question matters. The Guardian account says the anonymous account had more than 200,000 subscribers at one point. That is a historical audience figure, not a current total for an operating account, and it was not independently corroborated in the additional checks.⁠1

The earnings figure also needs correction. The initial Guardian account gave “¥500,000 to ¥75,000” per month. September 30 ORICON reporting instead gives 月50万～75万円: ¥500,000–¥750,000. Yomiuri’s May 23, 2026 report gives the same range. The erroneous upper figure is ten times too small; simply reversing the descending English range would produce a different, still incorrect estimate.⁠1⁠14⁠15

Those earnings remain alleged, not audited. Yomiuri said the estimate was based on information the uploader had published on another website. Neither that account nor ORICON’s description establishes verified payments or a judicial finding of monthly income. Multiplying the estimate across the posting window would not establish proven total revenue.⁠14⁠15

The older Yomiuri report described at least 188 videos posted from July 2024 until filing; the September 30 ORICON account specifies July 2024 through September 2025. The later report supplies the more specific posting window. Neither formulation establishes the exact lawsuit filing date.⁠14⁠15

The Japan Actors Union supported Tsuda’s action and thanked him for his courage, while acknowledging the emotional strain of the proceedings. Its organizational statement said it would review the formal judgment and subsequently comment on its website. No publication date was announced, and no verified appeal plan or case number is established in the available material.⁠23⁠18

The concrete outcome is substantial but limited: Tsuda’s removal objective was achieved, and voices were recognized as eligible for publicity-rights protection. Whether another disputed use satisfies the exclusive-commercial-exploitation test cannot be answered by that recognition alone.

Sources consulted include The Guardian⁠1, The Japan Times⁠2, Nippon.com⁠3, Tokyo Shimbun⁠4, Nikkei Asia⁠5, and Bengoshi.com reporting reproduced by NTT Docomo⁠6.

Additional coverage includes Anime News Network⁠7, Sanyo Shimbun⁠8, Voice Over Herald⁠9, Inven Global⁠10, Bengoshi.com reporting reproduced by Livedoor News⁠11, FandomWire⁠12, and WPNews⁠13.

Japanese reporting also includes ORICON via Web To-o⁠14, Yomiuri Shimbun⁠15, Nikkei⁠16, ITmedia NEWS⁠17, Sponichi⁠18, Jiji Press⁠19, and Fukushima Minyu⁠21.

Other consulted sources are The Straits Times⁠20, TV Asahi reporting reproduced by NTT Docomo⁠22, Nikkan Sports⁠23, AFP⁠24, TheJapanTimes.jp⁠25, 4Gamers⁠26, Mainichi Shimbun⁠27, and ORICON reporting reproduced by NTT Docomo⁠28.