Taylor Swift files for voice trademark protection: Copyright law cannot regulate AI voice cloning, trademark battles may fill the gap

Taylor Swift’s licensing company has filed three trademark applications with the USPTO, targeting her own voice and stage persona. Currently, copyright law has a structural gap regarding AI voice cloning, and trademark law may be the most effective tool available to fill it. (Context: How will gambling and prediction markets destroy the world? The darker side is yet to come) (Background: Trump announces "National AI Legislative Framework"! Pushing for a single federal regulation to firmly protect US AI hegemony) Copyright law protects songs, but not the voice itself, a legal loophole that makes it impossible to pursue AI voice cloning. Taylor Swift’s team chose a different path last week: filing three trademark applications with the USPTO to turn her voice and stage persona into legal weapons that can claim "likelihood of confusion." The applications were filed by her licensing company, TAS Rights Management, totaling three: two sound trademarks—audio clips of her saying "Hey, it’s Taylor Swift" and "Hey, it’s Taylor"—and one image trademark, featuring a photo of her on stage holding a pink guitar and wearing a colorful iridescent bodysuit. The audio clips come from a promotional campaign she did for her new album, *The Life of a Showgirl*, on Amazon Music. Application numbers sn99784980, sn99784979, and the image application sn99784977 are all publicly available. Swift’s team has not explicitly stated whether the trademarks are aimed at AI, but the legal community’s interpretation is unanimous: this is preparing for the AI era. Copyright protects creative works: song melodies, lyrics, and master recordings, but it does not protect "a person’s speaking voice itself." This means that as long as AI tools mimic the voiceprint rather than directly copying the recording, they can bypass copyright law. In 2024, Universal Music Group (UMG) issued a DMCA takedown notice for an AI-generated Drake-style song, but ultimately had to rely on producer Metro Boomin’s producer tag to claim copyright, rather than Drake’s voice itself. This is not an isolated case, but a blind spot that the copyright system did not anticipate when it was originally designed. IP lawyer Josh Gerben points out that the standard for trademark law is "likelihood of confusion," rather than the "substantial similarity" required by copyright law: this threshold is lower and better suited for dealing with AI voice cloning. In other words, even if the AI-generated voice is not a verbatim copy, it could constitute trademark infringement if it causes consumer confusion. Precedents already exist: actor Matthew McConaughey received USPTO approval for 8 trademarks in December 2025, including his signature line "Alright, alright, alright" and video clips, which the industry views as a direct precedent for stars using trademark law to combat AI impersonation. Of course, skepticism has also emerged. Northeastern University law professor Alexandra Roberts points out a core issue: the audio clip Swift applied for is a promotional message for Amazon Music, not an independently used identifier. The standards for traditional sound trademarks, such as the NBC chimes or the MGM lion roar, are all based on identifying a commercial source in a single, independent manner, rather than being attached to a specific promotional context. If the USPTO determines that this recording does not meet the requirements of an "identifying symbol," it may issue an initial refusal. However, Swift’s team could still supplement with samples that better meet the requirements, and the process is not over. UCLA law professor Xiyin Tang’s observation points to another dimension: the primary function of these trademarks may not be to win lawsuits, but to deter. Once a trademark is registered, potential infringers must assess legal risks before deciding to produce or distribute AI-cloned content. Even if the trademark is eventually challenged in litigation, this pre-emptive deterrent effect has already occurred. Swift’s path of legal patching comes against the backdrop of a larger legislative vacuum. In January 2024, AI-generated deepfake pornography of her went viral on X, with the hottest single post garnering over 45 million views. In August of the same year, Trump posted AI-generated images on Truth Social falsely implying "Swifties for Trump," and Swift publicly stated that AI deepfakes made her feel fearful. At the federal level, the NO FAKES Act, which aims to establish federal-level protection for likeness and voice rights, is still under congressional review, with support from SAG-AFTRA, UMG, Warner, and OpenAI. However, whether and when the bill will pass remains unknown.