Owning a trademark on your own voice means controlling both sides of the equation.
In late April, Taylor Swift's intellectual property entity quietly filed three trademarks with the U.S. Patent and Trademark Office — two capturing her voice, one her visual likeness — marking a deliberate legal maneuver against the growing threat of AI-generated impersonation. The move comes after Swift endured two deeply personal violations: nonconsensual deepfake pornography and a fabricated political endorsement. In reaching for trademark law where Right of Publicity protections fall short, Swift and others like Matthew McConaughey are asking an old legal system to hold the line against a new kind of identity theft — and in doing so, they may be writing the early chapters of how fame itself gets legally defined in the age of artificial intelligence.
- AI tools have already been weaponized against Swift twice — once through nonconsensual pornographic deepfakes, and once through a fabricated Trump endorsement video that spread without consequence.
- Existing Right of Publicity laws offer only partial shelter, leaving a dangerous gap that bad actors and AI systems can exploit with relative impunity.
- Swift's team filed two voice trademarks and one visual trademark through TAS Rights Management, creating a federal, actively enforceable legal claim over her identity.
- Matthew McConaughey made similar filings in January, signaling this is becoming a coordinated industry strategy rather than a one-off legal gambit.
- The same trademark that blocks unauthorized AI misuse could also empower celebrities to license synthetic versions of themselves on their own terms — making this both a shield and a business tool.
- The filings are real but untested against AI-specific claims in court, and the deeper question remains whether intellectual property law can evolve fast enough to match the technology it is being asked to restrain.
On a Friday in late April, three trademark applications arrived quietly at the U.S. Patent and Trademark Office, filed by TAS Rights Management — the entity that handles Taylor Swift's intellectual property. Two are sound trademarks, capturing Swift's voice saying "Hey, it's Taylor Swift" and "Hey, it's Taylor." The third is a visual trademark describing her in vivid stage detail: pink guitar, iridescent bodysuit, silver boots, purple lights. All three have been approved and are awaiting an examining attorney.
The filings drew little attention until intellectual property attorney Josh Gerben flagged them in a blog post, describing them as tools "specifically designed to protect Taylor from threats posed by artificial intelligence." The concern is well-founded. Right of Publicity laws give celebrities some recourse when their likeness is used commercially without consent, but those protections have limits. A registered federal trademark adds a harder, more enforceable layer of defense.
For Swift, this is not abstract. She has already been victimized twice in high-profile ways: nonconsensual AI-generated pornographic images of her circulated widely online, and during the 2024 presidential campaign, a fabricated video showed her appearing to endorse Donald Trump — a clip Trump himself reposted. Neither incident led to meaningful accountability.
Swift is not moving alone. In January, Matthew McConaughey secured eight trademarks, including one on his signature "Alright, alright, alright," with attorneys citing the same dual purpose: blocking unauthorized AI use while preserving the right to develop legitimate AI-driven opportunities on his own terms. That combination — defensive and commercial — reflects something important about the moment. The technology that enables abuse also enables new creative and economic possibilities, and owning a trademark on your own voice means controlling both sides.
These filings are early and untested against AI-specific claims in court. But as the gap widens between what existing law was built to protect and what celebrities now actually need protection from, more will almost certainly follow Swift and McConaughey's lead. The harder question is whether the legal infrastructure can keep pace with the technology it is being asked to restrain.
On a Friday in late April, three trademark applications landed quietly at the U.S. Patent and Trademark Office — and together they sketch the outline of a legal strategy that may define how celebrities fight back against artificial intelligence for years to come. The filer was TAS Rights Management, the entity that handles Taylor Swift's intellectual property. The applications have since been approved and are waiting to be assigned to an examining attorney.
Two of the filings are sound trademarks. One captures Swift's voice saying "Hey, it's Taylor Swift." The other captures her saying "Hey, it's Taylor." The third is a visual trademark, and its description reads like a stage photograph frozen in legal amber: Swift holding a pink guitar with a black strap, dressed in a multicolored iridescent bodysuit with silver boots, standing on a pink stage before a multicolored microphone, purple lights glowing behind her.
The filings went largely unnoticed until Monday, when intellectual property attorney Josh Gerben flagged them in a blog post. His read on their purpose was direct: the trademarks are, in his words, "specifically designed to protect Taylor from threats posed by artificial intelligence." The concern driving that assessment is one that has been building across the entertainment industry — that existing law may not be enough to stop AI systems from generating convincing, unauthorized versions of a celebrity's voice or face.
The legal landscape here is genuinely complicated. "Right of Publicity" laws already give celebrities some recourse when their image or likeness is used to sell products without permission. But those protections have limits, and trademark registration adds a separate, reinforced layer. A registered trademark is a federal claim, actively enforceable, with teeth that Right of Publicity suits don't always have.
For Swift, this isn't a theoretical concern. She has already lived through two high-profile incidents of AI-driven abuse. Pornographic deepfake images of her circulated widely online, making her the most visible victim of a problem that tech platforms and anti-abuse organizations have struggled to contain. Separately, during the 2024 presidential campaign, a fabricated video showed her appearing to endorse Donald Trump — and Trump himself reposted it as though it were genuine. Neither incident resulted in meaningful accountability for those responsible.
Swift is not alone in moving toward trademark law as a defensive tool. In January, actor Matthew McConaughey secured eight trademarks through the U.S. Patent and Trademark Office, including a sound trademark on his signature catchphrase, "Alright, alright, alright." Attorneys at the entertainment law firm Yorn Levine, which handled McConaughey's filings, told Variety the goal was twofold: to protect his voice and likeness from unauthorized AI use, and to preserve his ability to develop new opportunities using AI on his own terms.
That dual purpose — defensive and commercial — is worth sitting with. The same technology that enables a bad actor to generate a fake endorsement video can also enable a celebrity to license a synthetic version of themselves for legitimate projects. Owning a trademark on your own voice and image means controlling both sides of that equation.
Representatives for Swift did not respond to requests for comment, nor did Rebecca Liebowitz, the Venable partner listed as the attorney on the filings. The applications themselves do the talking for now.
What Swift and McConaughey are doing is early, but it is not isolated. As AI tools grow more capable and more accessible, the gap between what the law was designed to protect and what celebrities actually need protection from keeps widening. Trademark filings like these are one attempt to close it — imperfect, untested in court against AI-specific claims, but real. More celebrities will almost certainly follow. The question is whether the legal infrastructure will keep pace with the technology it is being asked to restrain.
Notable Quotes
The trademarks are specifically designed to protect Taylor from threats posed by artificial intelligence.— Josh Gerben, intellectual property attorney
The filings were made to protect McConaughey's voice and likeness from unauthorized AI use, and to protect him in developing new opportunities using AI.— Attorneys at Yorn Levine, as told to Variety
The Hearth Conversation Another angle on the story
Why trademarks specifically? Couldn't she just sue under existing law if someone misused her voice?
She could try, but Right of Publicity laws are state-level, inconsistent, and weren't written with AI in mind. A federal trademark is a harder, cleaner weapon.
What does it actually mean to trademark a phrase like "Hey, it's Taylor"?
It means that phrase, in her voice, in a commercial context, belongs to her legally. Anyone using it to sell something — or to impersonate her — has a problem.
The visual trademark description is oddly specific. Why describe the outfit and the stage lighting in such detail?
Precision is the point. The more specific the description, the more clearly defined the protected image. Vague trademarks are easier to work around.
She's already been victimized twice — the deepfakes, the fake Trump endorsement. Does this filing come too late?
For those incidents, yes. But trademark registration is about future enforcement. It changes the calculus for anyone thinking about doing it again.
McConaughey trademarked "Alright, alright, alright." That feels almost playful compared to what Swift is dealing with.
The stakes are different, but the legal logic is the same. Both are trying to own the thing that makes them recognizable before AI can commodify it without permission.
Is there something uncomfortable about a celebrity trademarking their own voice? Like — is that a new kind of monopoly on identity?
It's a real tension. But the alternative, in the current moment, is that someone else effectively owns it by being first to generate a convincing fake.
What happens next — does an examiner just rubber-stamp these, or is there real scrutiny?
There's a review process, and examiners can push back. But all three applications have already cleared the initial approval stage, so the path looks clear.