1. Introduction: The voice behind the character
A familiar voice can be just as recognisable as a face. Whether associated with a fictional character, public figure or someone close to us, the human voice carries distinctive qualities through which individuals are recognised and identified.
This relationship between voice and identity is particularly visible in Japan, where seiyū (voice actors) occupy a prominent place in the country’s cultural and creative industries. Far from being anonymous performers behind animated characters, Japanese voice actors often enjoy considerable recognition, with dedicated fan communities and careers extending across animation, video games, music and live performances. As Rayna Denison has explored, Japanese voice acting forms part of a distinctive transmedia star system in which the voice itself becomes an important element of performers’ public identities.
Advances in generative artificial intelligence (GenAI), however, have introduced new challenges. AI systems can produce convincing imitations of an individual’s voice without necessarily reproducing existing recordings, raising questions about the adequacy of copyright, performers’ rights and personality rights.
On 30 September 2026, the Tokyo District Court delivered a significant judgment concerning the alleged unauthorised AI imitation of Japanese voice actor Kenjiro Tsuda’s voice. Although dismissing the claim, the Court expressly recognised that a person’s voice may qualify for protection under Japanese publicity rights.
The decision comes amid growing demands for protection against AI voice cloning, including legislative initiatives in Ireland and Denmark and the United Kingdom’s Save Our Voices Now campaign. While these initiatives respond to similar concerns, they adopt different legal approaches and reflect potentially different justifications for protecting vocal identity. These differences may have important consequences as the technology and its uses continue to evolve.
2. Japan: Recognising vocal identity under publicity rights
The proceedings before the Tokyo District Court (30 September 2026, Reiwa 7 (Wa) No. 70635) concerned 188 videos published on TikTok between July 2024 and September 2025. The videos featured narration which Japanese actor and voice performer Kenjiro Tsuda alleged had been generated using AI to imitate his distinctive voice without authorisation.
Tsuda, recognised for voicing characters including Kento Nanami in Jujutsu Kaisen and Seto Kaiba in Yu-Gi-Oh!, sought removal of the videos on the basis of publicity rights and the Japanese Unfair Competition Prevention Act.
The case raised an important question: whether an individual’s voice, independently of their image or recorded performances, could constitute a legally protected aspect of identity under publicity-rights doctrine.
Japanese publicity rights have developed through case law, notably the Supreme Court’s 2012 Pink Lady judgment, which recognised protection against the unauthorised exploitation of the commercial attraction associated with an individual’s identifying characteristics, where the use is directed exclusively towards exploiting that attraction.
In September 2026, the Tokyo District Court applied this reasoning to the human voice, recognising that, like a person’s likeness, a voice can represent an individual’s personality. Accordingly, the unauthorised use of a voice possessing customer-attracting power may infringe publicity rights where the use is directed exclusively towards exploiting that commercial attraction.
The Court nevertheless dismissed Tsuda’s removal request because the disputed videos had already been deleted, leaving no further obligation on the platform to remove them. The Court’s recognition of voice as a potential subject of publicity-rights protection was therefore not necessary to the dismissal of the claim. The Court did not determine whether the particular narration infringed Tsuda’s rights.
The judgment is consequently significant for its recognition that vocal identity may attract publicity-rights protection, rather than for establishing a general prohibition on AI voice cloning.
The judgment also followed an important governmental initiative. On 7 August 2026, a study group established by Japan’s Ministry of Justice and chaired by Professor Yoshiyuki Tamura published a report examining civil liability for the unauthorised use of portraits, voices and other personal characteristics, subsequently updated on 22 October to incorporate the judgment. Rather than proposing new rights, the report clarifies how existing legal principles may apply to unauthorised digital imitation.
Importantly, the report demonstrates that Japanese protection of vocal identity is not confined to commercial interests. While publicity rights protect the economic value associated with identifying characteristics, the study group separately examines the right not to have one’s voice used without justification, grounded in the protection of personal interests. This distinction is particularly relevant to individuals whose voices lack demonstrable commercial attractiveness and illustrates how different justifications for protection may coexist within a single legal system.
Voice, culture and creative labour
The Japanese cultural context is particularly relevant. Voice acting has developed a distinctive relationship between performers, fictional characters and audiences, where the voice may simultaneously contribute to a character’s artistic identity, a performer’s professional identity and the commercial value of a production.
Concerns about AI-generated performances have prompted significant industry mobilisation. In October 2024, Japanese voice actors launched the NOMORE Unauthorised Generative AI campaign, drawing attention to the unauthorised reproduction and dissemination of their voices.
More recently, on 15 October 2026, the Japan Actors Union published guidelines on generative AI uses of voices, promoting principles concerning authorisation, control over voice data and appropriate remuneration.
Together, these initiatives demonstrate a response extending beyond judicial recognition, highlighting questions of consent, labour and professional autonomy. While seiyū culture should not be regarded as the legal basis for the judgment, it provides important context for understanding the growing attention to vocal identity in Japan.
3. Emerging approaches beyond Japan
Japan is not alone in confronting AI-generated vocal imitation. Across Europe, legislative proposals and industry campaigns demonstrate growing concern about the adequacy of existing protections, albeit through different legal approaches.
Ireland: Protection against misuse and harm
Ireland’s Protection of Voice and Image Bill 2025 proposes criminal offences addressing certain unauthorised uses of an individual’s name, photograph, voice or likeness, including in advertising, fundraising and political communications, where the specified conditions concerning intent to cause harm or recklessness are satisfied.
During its Second Stage debate in June 2026, concerns included AI-enabled fraud, manipulation and misuse of digital identity. The proposal therefore emphasises preventing particular harmful conduct rather than recognising a general proprietary right over one’s voice.
United Kingdom: Control and legal recognition
In the United Kingdom, the Save Our Voices Now campaign, has called for stronger statutory protection against unauthorised voice cloning. The UK government’s March 2026 Report on Copyright and Artificial Intelligence similarly acknowledges limitations in existing protections, noting that copyright generally protects works and recordings rather than distinctive vocal characteristics. It proposes exploring a new digital replica or personality right.
These developments raise questions about whether vocal identity should be protected as an economic interest capable of ownership or as an aspect of personhood deserving protection irrespective of commercial value.
Denmark: Copyright-related protection
Denmark’s Bill, introduced in October 2026, proposes amendments to copyright legislation protecting individuals against the unauthorised public dissemination of realistic digitally generated imitations of their personal characteristics, including voices. It also proposes specific protection for performers against imitations of their artistic performances, subject to limitations safeguarding freedom of expression.
The use of copyright law to protect personality-related interests nevertheless raises questions about the suitability of this regulatory framework, as explored by Francesca Portante D’Alessandro in her comparative analysis of the Danish and Dutch proposals.
These emerging approaches raise questions not only about the adequacy of existing legal protections, but also about which regulatory mechanisms are best suited to addressing AI voice cloning, how they might operate together, and whether they can respond effectively as the technology evolves.
4. Different rights, different philosophies?
Japan’s recent publicity-rights judgment foregrounds the commercial value of a person’s voice, while its broader legal framework also recognises personality-based interests. Ireland’s proposed legislation emphasises protection against particular forms of misuse and harm; the UK debate raises questions of individual control and legal recognition; and Denmark’s proposed copyright-related protections address both personal characteristics and artistic performances.
These different justifications are neither mutually exclusive nor confined to particular jurisdictions. Nevertheless, their relative emphasis may shape whose voices receive protection, which forms of imitation or exploitation are prohibited, and what remedies are available.
First, whose voices deserve protection?
A framework centred on commercial attractiveness may be particularly relevant to established performers and public figures. However, as Japan’s Ministry of Justice study group emphasises, publicity-rights protection does not necessarily require widespread fame: commercial attractiveness may also exist within particular communities or markets. Nevertheless, AI voice cloning can equally affect individuals whose voices possess little or no demonstrable commercial value.
This raises a fundamental question about whether protection should depend on the economic value of a person’s voice, or extend more broadly to the personal interests associated with vocal identity.
Second, what conduct should protection regulate?
Voice cloning involves distinct activities: collecting recordings, processing voice data, generating synthetic imitations and subsequently distributing or exploiting those outputs.
Copyright and neighbouring rights may protect recordings and performances, while privacy and data protection law may regulate the collection and processing of identifiable voice data. Personality and publicity rights may address different forms of unauthorised exploitation of vocal identity.
Yet permission to reproduce a recording does not necessarily authorise the exploitation of a synthetic imitation. Conversely, protection against commercial imitation may not address the preceding collection and processing of voice data.
The classification of voice as a performance, personal characteristic, economic asset or personal data therefore has consequences for the conduct falling within legal protection.
Third, what does it mean to own one’s voice?
Proprietary protection could strengthen individuals’ control over commercial exploitation, but also raises questions concerning licensing, transferability and bargaining power.
If vocal identity is principally treated as an economic asset, could performers be pressured into granting broad authorisations for future synthetic performances? Would such arrangements adequately recognise the continuing relationship between individuals and their voices?
A personality-based approach might better recognise that enduring connection, although it would present its own challenges concerning consent, permissible uses and remedies.
Finally, how should protection accommodate creative freedom?
Voice imitation is not inherently harmful. It has long featured in parody, satire and artistic performance, while synthetic voices may facilitate accessibility, translation and new forms of cultural production.
The challenge is therefore not simply prohibiting imitation, but determining when it constitutes unlawful appropriation, interference with personal autonomy or legally cognisable harm.
The underlying justification for protection may therefore influence where these boundaries are drawn.
5. Conclusion: Finding the right voice for legal protection
The Tokyo District Court’s September 2026 judgment represents an important development in recognising vocal identity under Japanese publicity rights. Yet significant questions remain concerning its application to AI-generated imitation and individuals without established commercial recognition.
Initiatives in Ireland, the United Kingdom and Denmark demonstrate that these questions extend beyond Japan. They also suggest that, despite growing recognition of the need to address unauthorised AI voice cloning, there is less consensus about the interests warranting protection and the most appropriate legal mechanisms for safeguarding them.
The distinction between voice as an economic asset, personal characteristic or creative performance is not merely conceptual. Different legal justifications may shape whose voices are protected, which forms of exploitation they can resist and what remedies are available. Equally important is whether existing and emerging protections can operate together coherently and be effectively enforced as technologies and uses evolve.
As AI increasingly enables the reproduction of recognisable human voices, the central question may therefore be not simply whether the law should protect one’s voice, but what, and whom, that protection is ultimately intended to serve.
Acknowledgement
I would like to thank Professor Tatsuhiro Ueno for organising the Deepfake and the Law Seminar at Waseda University on 3 September 2026, which I had the pleasure of attending. The discussions with Japanese and international scholars on the protection of voices and portraits, personality rights and creative freedom provided valuable inspiration for this piece.