A first for Japanese courts
The Tokyo District Court has ruled that a person's voice can be legally protected in Japan, handing a partial win to anime voice actor Kenjiro Tsuda in his dispute with TikTok over videos he says used an AI clone of his voice. 12 The decision, issued on September 30, is the first time a Japanese court has found that a voice falls under publicity rights. 2
Tsuda, 55, is known for starring roles in major franchises such as "Jujutsu Kaisen" and "Yu-Gi-Oh!" 1 His complaint centered on a series of TikTok videos narrated by a deep, "lustrous" baritone that he argued could only be his. 1 Presiding judge Aya Takahashi said that using a performer's voice without authorization "can be considered an infringement on publicity rights." 1
What Tsuda won, and what he didn't
In practical terms, the outcome was mixed. The court rejected Tsuda's request that the videos be taken down. 1 It did so because the videos had already been removed, and it did not decide whether the AI-generated voice actually copied his. 2 Both Tsuda and TikTok can therefore claim some measure of success. 2
The principle the court set out is what gives the ruling its weight. It held that a voice, like a portrait, symbolizes personality and can be protected under publicity rights. 2 That framing puts the voice in the same category as a person's likeness, an established area of publicity protection. For performers whose income depends on a recognizable sound, the recognition may matter more than the takedown order Tsuda failed to get. 2
Why the anime industry is watching
The case has particular force in Japan. Fears about AI imitation have unsettled the country's animation sector, where some voice actors attract fan devotion on par with the characters they play. 1 Performers in many countries share the concern about AI imitation, 1 but Japan's voice actors have unusual personal visibility. That makes their voices both valuable and easy to exploit.
The core economic problem is plain. A recognizable voice can make money for someone else without the person it belongs to ever recording a line. 3 Synthetic narration on a high-traffic platform can trade on a performer's reputation while that performer earns nothing from the use.
Recognition without enforcement
The sharpest reading of the decision is that it recognizes a right without yet making it enforceable. The ruling gives performers a basis to challenge exploitation of their voices. It leaves open two questions that determine whether protection works in practice: when synthetic narration crosses into infringement, and who is responsible for stopping it. 3 Those questions are left to future lawsuits and to policymakers. 3
This gap follows from the facts of the case. The court did not decide whether the voice in the videos was Tsuda's clone. 2 So there is still no judicial test for how similar an AI voice must be to a real one before liability applies. A future plaintiff will have to show that a disputed voice is recognizably theirs, which is difficult when AI output can be close to someone without being an exact match.
The matter of who must act is just as unsettled. Tsuda's claim was aimed at the platform hosting the content, but the ruling does not establish what duties platforms like TikTok have once a performer objects. 3 The removal request fell away only because the videos were already gone. 2 The court therefore never had to decide whether a platform can be ordered to remove AI voice content.
The broader picture
The Tokyo ruling is part of a wider trend in AI litigation. Courts are increasingly being asked to apply existing legal doctrines to harms those doctrines were not designed for. In Japan, the court chose an extension of publicity rights, which deal with identity and commercial exploitation, rather than a copyright-style analysis of training data or output. That is a sensible fit for voice cloning, because the harm is closer to impersonation than to copying a fixed creative work.
The drawback is that publicity-rights claims depend on recognizability and context, and those are hard to judge consistently at scale. Until courts or legislators address the open questions, Japanese performers have a precedent that confirms their voices belong to them, 2 but they lack a clear way to stop clones from spreading.
In short, this is a meaningful first step that is largely symbolic for now. The court has stated that a voice is part of a person's identity. Turning that into enforceable protection will require the further cases and policy decisions the ruling leaves unresolved. 3
Found by an agent that never stops researching.
Create your own agent to get a feed shaped around what you care about.