When Kenjiro Tsuda—the celebrated Japanese voice actor whose rich baritone has brought some of anime’s most iconic characters to life—first heard the unauthorized account, he immediately recognized the problem: a synthetic version of his voice was narrating videos about urban legends and conspiracy theories to an audience of more than 200,000 subscribers, and he hadn’t said a single word.
The resulting lawsuit, now reaching verdict in a Tokyo district court, may be the most consequential AI legal case you’ve never heard of—a watershed moment for voice actors, content creators, and the entire ecosystem of generative AI that is reshaping how human identity can be exploited, replicated, and monetized without consent.
The Voice That Built a Career
Tsuda, 55, spent decades training his instrument. He is perhaps most recognizable internationally as Nanami Kento in *Jujutsu Kaisen*, the globally dominant anime franchise, and as Seto Kaiba in *Yu-Gi-Oh!*—characters whose gravitas relies entirely on the distinctive weight and texture of a human voice shaped through years of disciplined craft.
In early 2026, Tsuda’s management team flagged an anonymous TikTok account that had posted more than 180 videos. The account used imagery stylistically reminiscent of a character Tsuda had voiced. More strikingly, each video featured a synthesized narration voice bearing an uncanny resemblance to Tsuda’s own deep baritone—melodically particular, with inflection patterns that fans of his work recognized immediately.
The account had accumulated over 200,000 followers and, according to filings from Tsuda’s legal team, generated approximately ¥500,000 (roughly $3,200 USD) in monthly revenue through TikTok’s creator monetization program.
Tsuda’s team requested removal. TikTok declined, classifying the voice as “generic male” and the similarity as “subjective.” The case proceeded to litigation.
Japan’s Legal Framework — And Its Gaps
Japan does not have an explicit statute protecting a person’s voice from AI replication. What it does have is a body of publicity rights law—protections governing how an individual’s name, likeness, and image may be used for commercial purposes without consent—and portrait rights inherited from decades of case law around photograph and image use.
Tsuda’s legal team has argued these frameworks apply equally to voice: that a recognizable voice, especially one closely associated with a performer’s commercial identity, constitutes a protected element of publicity rights. The Tokyo District Court is being asked to extend these precedents explicitly to the domain of AI-synthesized audio.
Japan’s justice ministry issued non-binding guidance earlier this year acknowledging that voices should be considered protectable under existing law—a move interpreted by legal observers as preparing the ground for exactly this kind of ruling. But guidance is not statute, and the court’s decision this week will carry weight that ministerial memos cannot.
“Our Industry Can Be Crushed”
The Japan Actors Union has rallied behind Tsuda under a campaign it calls “No More”—a direct appeal against unauthorized AI voice mimicry of working performers.
“A voice is the result of years of rigorous training,” said Yuko Sasaki, the union’s executive director. “The ability to control how it is used is a fundamental right.”
Fellow voice actor Bin Shimada, known for his long career in anime dubbing, was more blunt in his warning: “Our industry can be crushed if this is allowed to continue.”
The concern is not hypothetical. AI voice cloning technology has advanced dramatically in the past two years. Tools available for as little as $10 per month can now generate high-quality synthetic speech that mimics a specific individual’s voice from a sample as brief as thirty seconds. Some platforms require no sample at all—users simply select from pre-trained celebrity or performer profiles scraped from publicly available recordings.
The economic implications for voice acting are severe. A profession that depends on the uniqueness of individual vocal identities faces a potential race to the bottom: why pay a voice actor for a campaign if an AI can replicate their signature sound for a fraction of the cost? The Tsuda case makes this abstract threat vivid and legally contested.
TikTok’s Defense and the Platform Liability Question
TikTok’s legal position revolves around two arguments. First, the platform characterizes the disputed voice as unremarkably generic—a position that strains credibility given the specificity with which fans and colleagues identified the resemblance. Second, and more substantively, TikTok argues it is merely a platform hosting user-generated content and bears no special responsibility for the creative choices users make within that content.
This second argument echoes the safe harbor defenses that have shielded platforms from copyright liability for user uploads in the United States under Section 230 and the Digital Millennium Copyright Act. Japan’s legal landscape does not have identical protections, but the principle—that platforms are not publishers—has carried weight in Japanese courts when applied to text and image content.
Whether it applies to AI-synthesized audio specifically modeled on an identifiable individual is precisely the question the Tokyo District Court is now weighing.
Intellectual property professor Tatsuhiro Ueno, who has followed the case closely, told reporters he expects the verdict to attract significant international attention. “The question of whether a performer’s voice is protectable against AI cloning will be decided in courts around the world in the coming years,” he said. “Japan may be first.”
A Global Pattern
Tsuda’s case does not exist in isolation. In 2024, OpenAI issued a public apology to actress Scarlett Johansson after a synthetic voice in its ChatGPT product bore a striking resemblance to her distinctive sound—an incident that prompted urgent conversations about consent in voice AI development. Musicians, including prominent artists in the United States and South Korea, have pursued similar disputes with platforms distributing AI-generated music that replicates their vocal style.
In the European Union, the AI Act—now entering its enforcement phase—includes specific provisions on high-risk AI applications and requires transparency disclosures when synthetic media involves realistic representation of real individuals. The United States has no equivalent federal law, though multiple states have passed or are considering legislation addressing AI-generated likeness and voice.
The common thread in each dispute is a gap between the pace of AI capability and the frameworks society has developed to govern its use. Voice cloning is now a commodity technology, accessible and affordable. The legal systems attempting to constrain its misuse are still catching up.
What Comes Next
The Tokyo District Court’s ruling—expected imminently—will not resolve the global debate. But it will contribute to a growing body of precedent that regulators, platforms, and performers around the world will cite in the cases that inevitably follow.
If the court rules in Tsuda’s favor and orders TikTok to remove the content, it will affirm that AI-synthesized voice replication falls within the scope of existing publicity rights protections—a finding that will bolster voice actors and performers negotiating new protections into their contracts and arguing for legislative action.
If TikTok prevails on a platform-liability defense, it will signal that AI-cloned voices can circulate freely on major platforms absent explicit statutory intervention—a finding that will accelerate legislative efforts in Japan and provide an urgent reference point for policymakers elsewhere.
Either way, the question Tsuda has put before the court will not disappear: in an age when a machine can replicate the instrument that defines a performer’s identity and livelihood, does the law protect what a person trains for decades to become?
The voice actor who gave gravity to Nanami Kento—stoic, precise, unwilling to let injustice pass—is now asking that question in real life. The answer matters far beyond anime.
*Word count: ~1,020 words*
