Taylor Swift’s New Trademarks: A Preemptive Strike on AI Deepfakes
Taylor Swift has quietly moved her AI worries into the legal arena. Her company, TAS Rights Management, recently filed three trademark applications with the U.S. Patent & Trademark Office covering both sound and image. Two applications focus on short audio clips in which she introduces herself—“Hey, it’s Taylor Swift…” and “Hey, it’s Taylor…”—and promotes her album “The Life of a Showgirl” on services like Amazon Music Unlimited and Spotify. A third filing covers an image of Swift performing on stage in a glittering bodysuit, playing a pink guitar during The Eras Tour. These moves are directly tied to the rise of AI deepfakes and synthetic media that can convincingly clone a celebrity’s voice or face. Instead of waiting for a crisis, Swift is creating a legal record of what her authentic voice and likeness sound and look like, so she can challenge AI-generated imitations more effectively.
How AI Deepfakes Create ‘New’ Songs, Videos and Statements
AI tools now make it possible to generate Taylor Swift deepfake clips in minutes, using just a brief sample of her voice or a handful of photos. Voice-cloning models can synthesize speech that sounds uncannily like a star, allowing bad actors to produce AI fake songs, fake brand endorsements or even fabricated political statements. Video tools can map a celebrity’s face onto someone else’s body, creating synthetic performances or explicit content that never actually happened. Swift has already felt the impact: a major platform temporarily blocked searches for her name after explicit AI images spread, and AI-generated visuals falsely suggested she backed a political campaign. As deepfakes become more realistic and easier to share, the reputational damage grows—fans can be misled, brands can be tricked and misinformation can travel faster than any correction. For artists, that means their most valuable asset—their identity—can be hijacked at scale.
Trademarks, Publicity Rights and the Rise of the Celebrity Voice Trademark
Swift’s filings are part of a broader push toward the celebrity voice trademark as a shield against AI misuse. Traditionally, performers relied on copyright for recordings and state “right of publicity” laws to stop unauthorized ads or merchandise. But AI can generate completely new content that mimics an artist without copying any existing track or photo, slipping through those older protections. Trademark law offers a different tool: by registering distinctive phrases and images, a star can argue that confusingly similar AI content misleads the public about endorsement or origin. Lawyer Josh Gerben notes that registering a voice as a mark is still largely untested, but it could fill the legal gap AI has opened. Several U.S. states are also tightening rules on AI misuse, and Tennessee’s ELVIS Act is a notable attempt to protect performers’ voices and likenesses more broadly in the music industry AI era.
From Matthew McConaughey to Scarlett Johansson: A Wider Creative Backlash
Taylor Swift is not alone in trying to lock down her identity. Actor Matthew McConaughey has pursued similar protection, filing trademarks for both his voice and image. His applications include iconic phrases such as “alright, alright, alright!” from the film Dazed and Confused, along with other short audio snippets and visual likeness rights, explicitly framed as a response to AI systems that could copy his persona. Other creatives are turning to the courts: Scarlett Johansson sued the app Lisa AI after it allegedly created an AI avatar in her likeness for an advertisement without consent. Meanwhile, more than 700 artists and creators have backed campaigns calling for stronger safeguards against AI deepfakes. Together, these moves signal a shift: voice, face and catchphrases are no longer just branding tools, but legal battlegrounds. High‑profile figures are setting precedents that lesser‑known performers may eventually rely on.
What It Means for Fans, Platforms and the Future of Music
As pop star image rights and voice trademarks expand, the ripple effects will hit fans, platforms and labels alike. For listeners, AI fake songs that convincingly sound like their favorite artist could flood social media and streaming sites, blurring what is official and what is synthetic fan fiction. Music platforms and networks such as YouTube and X are already under pressure to detect and remove deepfakes more quickly, and some are deploying tools specifically aimed at synthetic media. Future policies may require clearer labeling of AI‑generated content, stricter takedown procedures and cooperation with rights holders using trademarks and publicity laws. For the wider music industry AI landscape, this raises tough questions: Can fans still create playful remixes using AI? Where is the line between homage and infringement? Taylor Swift’s latest filings suggest that the default answer is shifting toward tighter control, and platforms will be expected to keep up.