Taylor Swift has filed with the USPTO to trademark her voice, joining a growing wave of performers seeking protection as AI-generated clones become faster and harder to police.
Taylor Swift has taken a step that many performers are increasingly considering: trying to lock down ownership of her own voice as AI clones spread.
According to filings made with the US intellectual property office, the singer submitted two sound recordings to the United States Patent and Trademark Office (USPTO).. Each recording opens with the phrase “Hey, it’s Taylor” and then includes an announcement tied to her album The Life of a Showgirl, which was released in early October.. A separate document filed on April 24 includes a photo of Swift performing on stage..
The filings were first spotted by intellectual property attorney Josh Gerben, and the application materials, as described, do not spell out additional details about the broader scope of what she is seeking.. When contacted for comment through Misryoum’s reporting process, Swift’s publicist did not immediately respond.
The move comes amid a shift in how quickly AI systems can imitate voices.. Advances in voice synthesis now allow a performer’s sound to be generated in seconds from a short clip—something that, just a few years ago, typically required longer recordings and more time.. For artists, that speed changes the practical risk: misuse can happen faster, at scale, and with less friction.
This strategy also mirrors an approach already tested in the US entertainment industry.. Actor Matthew McConaughey, in recent years, pursued trademark filings framed around protecting his voice from unauthorised use by AI models.. His applications have referenced recordings of him speaking phrases from his work, including a line associated with the 1993 film Dazed and Confused.. The underlying logic is the same: if AI can replicate a signature voice, trademark protection may help establish clearer boundaries for commercial or platform use.
For performers, the concern is not just casual imitation.. Many artists worry about the unchecked use of their image and voice by AI platforms—whether it’s aimed at advertising, impersonation, or other commercial activity.. “Ownership” here can be complicated, because voice likeness sits at the intersection of intellectual property, publicity rights, and consumer confusion.. Trademark law, in particular, is designed to help prevent the public from being misled about who is associated with a sound or brand.
The legal landscape in the United States is uneven.. Several states have passed laws that prohibit certain uses of voice and image through AI, but many of those laws tend to focus on malicious or explicitly commercial exploitation.. Broader protections exist in only a few places; one example is the ELVIS Act passed by Tennessee’s state legislature in 2024.. Even then, enforcement can vary, and legal remedies may depend on how a case is framed.
So far, relatively few performers have taken disputes into court.. One of the most closely watched examples involved Scarlett Johansson, who sued over an AI avatar in 2023 that used her likeness without consent in an advertisement tied to the app Lisa AI.. That case underscored a key point for artists: if protections are unclear or underused, litigation may become the only way to push platforms and developers to take consent seriously.
What makes Swift’s filing stand out is timing and visibility.. Swift’s name and catalogue give any dispute extra attention, and a trademark application can function as both a legal move and a signal.. If similar filings become more common, it may pressure platforms to adjust how they handle voice requests, verifications, and takedown processes—even when the underlying content is generated quickly and cheaply.
In practical terms, consumers also feel the impact.. As voice cloning improves, it becomes easier for people to be fooled by audio that sounds convincing, especially when it’s paired with familiar branding or release announcements.. The strongest defenses may come from a combination of legal tools—like trademarks and state laws—and platform policies that make misuse harder to monetize.