In a sign of how rapidly the battle over identity and artificial intelligence is escalating, pop icon Taylor Swift has taken a novel legal step: seeking trademark protection not just over her image, but over her voice itself.
Applications filed with the U.S. Patent and Trademark Office cover two audio clips featuring Swift’s spoken voice, along with a stage photograph showing the singer performing with a pink guitar. The filings were made through TAS Rights Management and are widely seen as a strategic response to the surge of AI-generated deepfakes mimicking celebrities.
The move pushes beyond traditional celebrity image rights into less-charted territory. While performers have long leaned on copyright and publicity laws to combat misuse, AI voice cloning has complicated those protections. Synthetic replicas can imitate an artist’s voice without directly copying any original recording, exposing a legal gray zone.
That gap appears to be what Swift’s camp is targeting.
One filing covers a promotional clip tied to her album The Life of a Showgirl and Amazon Music Unlimited. Another centers on a Spotify presave message for the same project. The performance image included in the applications appears aimed at shielding not just photographs, but recognizable visual markers associated with Swift’s persona.
The filing arrives amid growing alarm over AI-generated fakes involving public figures. Swift has repeatedly been caught in that storm — from fabricated endorsements and misleading advertisements to manipulated explicit imagery circulating online.
Legal observers say trademarks may offer a different weapon than copyright, particularly when confronting synthetic content designed to evoke a person without technically reproducing protected material. In that sense, the filings could be less about merchandise or branding in the traditional sense and more about building a defensive moat around identity itself.
The strategy echoes broader anxieties rippling through entertainment. Matthew McConaughey has pursued similar protections, part of a growing push among artists to establish ownership boundaries before AI imitation outpaces regulation.
What makes Swift’s move notable is its focus on spoken voice as a trademark asset — an approach still largely untested. If successful, it could signal a new template for musicians, actors and public figures confronting an era where likeness can be manufactured at scale.
The fight over AI deepfakes is often framed as a technology debate. Swift’s filing suggests it may increasingly become a battle over intellectual property — and over who owns a human identity when machines can copy its echoes.
