Tokyo court rules AI-cloned voice violates actor's rights

A Tokyo court has ruled that the human voice is protected under Japanese law, siding in part with anime voice actor Kenjiro Tsuda in the country's first case over an AI voice clone.

The dispute centered on a TikTok account that Tsuda said used AI to clone his "lustrous" baritone delivery. The anonymous account uploaded videos with AI-generated narration between July 2024 and the second half of last year. At its peak, it counted more than 200,000 followers. The court found that using an actor's voice without permission constitutes an infringement of their publicity rights, as reported by AFP. No Japanese court had previously ruled on the question.

Tsuda is one of Japan's most recognizable voice actors. Across a career spanning more than 30 years, he has starred in a large number of movies, series and video games, with his turn as Seto Kaiba in Yu-Gi-Oh! arguably his most iconic role. That prominence sat at the center of his complaint: the actor's legal team argued the account owed its popularity to the novelty of its "dubious, sordid content" being narrated by a prominent star.

The tell, according to Tsuda, was the unmistakably deep and "lustrous" quality of the AI-generated narration. TikTok, for its part, disputed the identification. In court records seen by AFP, the company argued the narration used a "generic male voice" whose similarity to Tsuda's was debatable.

The ruling was partial. The court declined Tsuda's request to order TikTok to remove the videos at issue, because the offending account had already been deleted. The decision on the underlying question — that a cloned voice can infringe publicity rights — stands regardless.

Why the ruling matters

The stakes run well beyond one actor and one TikTok account. Japan's national strategy has been notably pro-AI, and voice cloning tools have become cheap, fast and convincing. That combination leaves performers whose voices are their primary asset with few established legal protections. This case establishes, for the first time in Japan, that a voice itself — not just a name or image — can fall under publicity rights when reproduced by AI without consent.

For the anime industry, the ruling lands amid growing unease. Voice actors have raised concerns that more needs to be done to protect creators whose work, or likeness, is routinely ingested by AI systems. Last year, Japan's government asked OpenAI to ensure that anime and manga content it described as "irreplaceable treasures" was not being used to train the company's Sora 2 video generator in a way that violates copyright.

The two developments sketch the emerging fault line in Japan's AI policy: a government eager to promote the technology while fielding calls to shield the creative industries that form a significant part of the country's cultural exports.

An imperfect but significant win

Tsuda's victory is real but incomplete. The court affirmed the legal principle he sought to establish — that his voice is protected — but granted no removal order, since the anonymous account no longer exists. That gap points to the practical limits of enforcement: by the time a case reaches a courtroom, the cloned content may already be gone, and the person responsible may remain unknown.

The ruling nevertheless gives performers a precedent to invoke. Future plaintiffs in Japan can point to a court decision holding that unauthorized AI reproduction of a voice infringes publicity rights, rather than arguing the question from scratch.

It also raises questions platforms will have to answer. TikTok's defense — that the narration was a generic male voice — did not persuade the court on the central issue. How platforms assess and respond to voice-cloning complaints before cases reach litigation remains unresolved.

The decision comes as governments elsewhere grapple with similar questions over voice, likeness and AI-generated imitation. Japan's answer, delivered in a Tokyo courtroom over a TikTok narrator, now forms part of that global record.