Pop superstar Taylor Swift has filed trademark applications with the US Patent and Trademark Office for two audio clips and one photograph of herself, in what legal experts describe as a pioneering effort to protect her voice and likeness from artificial intelligence-generated deepfakes. The applications, filed on 24 April and owned by Swift's company TAS Rights Management, represent a growing trend among high-profile celebrities seeking legal tools beyond traditional copyright to defend themselves in the AI era.
In the two audio clips submitted, Swift is heard speaking promotional lines for her latest album, The Life of a Showgirl. The photograph shows her performing on stage during her record-breaking Eras Tour, dressed in a sequined outfit and holding a pink guitar. Trademark attorney Josh Gerben, who first publicised the filings, explained that the applications "are specifically designed to protect Taylor from threats posed by artificial intelligence." Swift is not alone in pursuing this approach: actor Matthew McConaughey became the first celebrity to successfully register such a trademark earlier this year, telling the Wall Street Journal that he wanted to "create a clear perimeter around ownership with consent and attribution the norm in an AI world."
The legal rationale centres on a gap that AI technology has opened in existing law. Copyright has long protected recorded music and performances, but AI tools can now generate entirely new content that mimics an artist's voice without copying any existing recording — meaning copyright law offers limited recourse. Trademark registration, Gerben argues, could provide "an additional layer of protection" alongside existing right-of-publicity statutes. Whether that argument holds up in court remains untested; no rulings have yet confirmed the approach, and the US Patent and Trademark Office warns that processing times for such applications can stretch to ten months.
The stakes are significant. Swift's image and voice have already appeared in a wide range of unauthorised AI-generated content, including fake advertisements, explicit imagery, and fabricated political endorsements — among them AI-manipulated images shared during Donald Trump's 2024 presidential campaign that falsely suggested Swift and her fans were endorsing him. Swift publicly supported Democratic candidate Kamala Harris during that election. The broader concern extends well beyond Swift: in 2024, actress Scarlett Johansson pressured OpenAI to withdraw a ChatGPT voice called "Sky" after she alleged it had been deliberately modelled on her own voice without consent, a dispute that ended without litigation but highlighted how ill-equipped current legal frameworks are to handle AI mimicry.
Swift's trademark move signals that celebrities and their legal teams are increasingly unwilling to wait for legislators to act, instead using available tools creatively to build protective legal barriers. Whether the strategy proves durable in court will likely depend on future test cases — and Swift's filing may itself become a landmark in that evolving legal landscape.