A voice that distinctive is also a voice worth faking.
In an age when artificial intelligence can conjure a convincing replica of any human voice or face within seconds, Taylor Swift has turned to the U.S. Patent and Trademark Office with a deceptively simple claim: that her voice and her image belong to her. The filings — modest in scope, deliberate in intent — represent a broader civilizational reckoning with what it means to own oneself in a world where identity has become raw material. Swift joins a growing chorus of public figures, from Matthew McConaughey in Hollywood to Amitabh Bachchan in Mumbai, all pressing the same urgent question: in the age of AI, who holds the rights to a human being?
- AI deepfakes have already weaponized Swift's likeness without her consent — a fabricated image falsely depicting her endorsing a presidential candidate was shared by the candidate himself during the 2024 U.S. election.
- Existing legal frameworks — copyright, right-of-publicity laws — were written for an era before a convincing human replica could be generated in seconds, leaving celebrities dangerously exposed.
- Swift's trademark filings covering two voice clips and an onstage photograph are a calculated legal escalation, adding a new layer of enforceable ownership on top of protections that have already proven insufficient.
- Matthew McConaughey is pursuing the same strategy, framing it not as a complaint but as the construction of a perimeter — a demand that consent and attribution become the baseline, not the exception.
- The fight is global: Indian stars including Bachchan, Anil Kapoor, and Allu Arjun have already sought court protections, signaling that the race to legally anchor personal identity is accelerating across industries and borders.
Taylor Swift has filed applications with the U.S. Patent and Trademark Office seeking recognition of something most would consider self-evident — that her voice and her face belong to her. The filings cover two short promotional audio recordings and a photograph of her performing onstage in a sequined outfit with a pink guitar. Modest in scope, they are pointed in intent: a legal stake driven into the ground at a moment when AI has made it disturbingly easy to put words in someone's mouth.
The threat is not hypothetical for Swift. During the 2024 presidential campaign, an AI-generated image falsely depicted her endorsing Donald Trump — and was shared by Trump himself. The episode illustrated how swiftly a celebrity's likeness can be weaponized in contexts they never chose and would never have sanctioned.
Trademark attorney Josh Gerben notes that the filings are designed to build a legal layer atop protections that already exist but were never designed for this moment. Copyright covers recorded music. Right-of-publicity laws protect commercial use of identity. But trademark law applied to a voice or a face creates a different kind of claim — one with additional teeth in commercial contexts where unauthorized AI use is most likely to cause harm.
Swift is neither the first nor the last to pursue this path. Matthew McConaughey has filed similar applications, telling the Wall Street Journal he wants a perimeter, not merely a grievance. In India, where the entertainment industry is enormous and digital identity law is still developing, Amitabh Bachchan has appeared before the Delhi High Court over unauthorized use of his name and voice, Anil Kapoor secured a sweeping injunction covering even his signature catchphrase, and singers and actors from Kumar Sanu to Allu Arjun have confronted comparable violations.
What unites these cases is a shared recognition that the law, however well-intentioned, was not built for a world where a convincing human replica can be conjured in seconds. The trademark strategy is partly symbolic — a declaration of ownership — and partly a practical expansion of legal options. Whether it holds will depend on how courts interpret these filings and how aggressively those filing them choose to fight. For now, the direction is clear: more celebrities, more applications, more litigation, and a question that can no longer be deferred.
Taylor Swift has asked the U.S. Patent and Trademark Office to recognize something most people would consider obvious: that her voice and her face belong to her. The applications, filed in recent weeks, cover two short audio recordings of the singer speaking and a photograph of her performing onstage in a sequined outfit while holding a pink guitar. The filings are modest in scope but pointed in intent — a legal stake in the ground at a moment when artificial intelligence has made it disturbingly easy to put words in someone's mouth.
The audio clips are promotional in nature. In one, Swift invites listeners to stream her album 'The Life of a Showgirl' on Amazon Music Unlimited. In the other, she encourages fans to presave the same record on Spotify ahead of an October 3 release. They are the kind of recordings that circulate constantly in the digital ecosystem — brief, casual, instantly recognizable. That recognizability is precisely the problem. A voice that distinctive is also a voice worth faking.
Swift's exposure to this threat is not hypothetical. During the 2024 U.S. presidential campaign, an AI-generated image falsely depicting her as an endorser of Donald Trump was shared by the former and now current president himself. The incident drew widespread attention and underscored how quickly a celebrity's likeness can be weaponized, with or without their knowledge, in contexts they would never have chosen.
Trademark attorney Josh Gerben, who has analyzed the filings, describes them as an attempt to build a legal layer on top of the protections that already exist. Copyright covers recorded music. Right-of-publicity laws, which vary by state, protect a person's image and identity from commercial exploitation. But those frameworks were written for a different era. Trademark law, applied to a voice or a face, creates a different kind of claim — one that could give Swift's legal team additional tools to pursue unauthorized uses, particularly in commercial contexts.
She is not the first to try this approach, and she will not be the last. Actor Matthew McConaughey has filed similar applications in recent months. Speaking to the Wall Street Journal, McConaughey framed the effort in terms of establishing clear ownership in an AI-driven world — one where consent and attribution should be the baseline expectation, not the exception. His language was deliberate: he wants a perimeter, not just a complaint.
The problem extends well beyond Hollywood. In India, where the entertainment industry is vast and the legal infrastructure around digital identity is still catching up, several prominent figures have already sought protection through the courts. Amitabh Bachchan appeared before the Delhi High Court to challenge the unauthorized commercial use of his name, voice, and image. Anil Kapoor secured a broad injunction in 2023 that covered not just his likeness but his signature catchphrase, 'Jhakaas.' Singers Kumar Sanu and Asha Bhosle, along with actor Allu Arjun, have faced similar unauthorized uses of their identities in AI-generated content.
What connects all of these cases is a recognition that existing law, however well-intentioned, was not designed for a world in which a convincing replica of a person can be generated in seconds. The trademark strategy is partly symbolic — a declaration of ownership — and partly practical, adding another legal avenue to pursue bad actors. Whether it will prove effective depends on how courts interpret these filings and how aggressively rights holders choose to enforce them.
For now, the filings signal where this is heading. More celebrities, more applications, more litigation. The question of who owns a person's voice and face in the age of AI is no longer abstract. Swift, McConaughey, Bachchan, and Kapoor are all, in their different ways, insisting on an answer.
Citações Notáveis
We want to create a clear perimeter around ownership, with consent and attribution the norm in an AI world.— Matthew McConaughey, speaking to the Wall Street Journal
The filings are aimed at creating an additional layer of protection beyond existing laws.— Trademark attorney Josh Gerben, on Swift's USPTO applications
A Conversa do Hearth Outra perspectiva sobre a história
Why trademark a voice clip instead of just relying on existing laws?
Because existing laws have gaps. Copyright protects recordings, publicity rights protect commercial use of your image — but neither was built with AI impersonation in mind. Trademark adds another angle of attack.
What does a trademark on a voice actually give you that you didn't have before?
A different legal hook. If someone uses a trademarked voice clip to sell a product or mislead consumers, you can pursue them under trademark law, which has its own remedies and its own teeth.
The clips Swift filed are promotional — album ads, essentially. Is that significant?
It's strategic. Commercial speech is exactly where trademark law is strongest. She's not trying to trademark every word she's ever spoken — she's planting a flag in the territory most likely to be exploited.
The Trump deepfake during the 2024 campaign — did anything come of that legally?
Not in any significant public way. Which is part of the point. The existing tools weren't enough to stop it or punish it effectively. That's the gap these filings are trying to close.
McConaughey's language about a 'perimeter' is interesting. Is that how lawyers think about this?
It's how anyone thinks about it when the threat is diffuse. You can't sue every bad actor individually. You build a legal structure that makes unauthorized use clearly, unambiguously wrong — and then you enforce selectively but visibly.
The Indian cases feel different in character — court injunctions rather than trademark filings. Why the difference?
Different legal systems, different tools available. India's courts have been willing to grant broad injunctions protecting celebrity identity. The U.S. system pushes more toward proactive registration. Same goal, different path.
Is any of this actually going to stop deepfakes?
Probably not stop them. But it changes the calculus for anyone thinking about commercial exploitation. Legal risk is a real deterrent, even if it's not a perfect one.