Taylor Swift is moving to lock down her voice and image through federal trademark law, following a growing trend among high-profile celebrities seeking legal tools against AI-generated content that mimics their likeness without consent.
As first spotted by IP attorney Josh Gerben and reported by Variety, Swift's management entity TAS Rights Management filed a trio of trademark applications with the U.S. Patent and Trademark Office late last week. The filings cover different elements of Swift's identity, each targeting a specific surface area where AI-generated imitation poses commercial and reputational risk.
The 3 Trademarks | Voice Phrases and a Photograph
Two of the three filings are voice-related applications. They cover Swift speaking the phrases "Hey, it's Taylor Swift" and "Hey, it's Taylor" in her own voice. These short identity phrases are the kind most commonly cloned by AI voice synthesis tools, which can replicate a celebrity's vocal characteristics from a relatively small audio sample and generate convincing fake clips at scale.
The third application covers a specific image, described as a photograph of Taylor Swift holding a microphone, representing a recognizable visual identity marker tied directly to her Eras Tour persona and broader public image.
Together, the three trademarks establish a legal baseline Swift's team can use to challenge unauthorized commercial use of AI-generated content that mimics her voice or replicates her documented visual presentation.
Following McConaughey | The Celebrity AI Trademark Playbook
Swift is not the first major celebrity to pursue this route. Actor Matthew McConaughey filed a similar trademark application covering his voice earlier this year, establishing the phrase "alright, alright, alright" as a protectable commercial identifier. That filing drew widespread attention from IP attorneys as an early signal of how celebrities and their legal teams are adapting to the AI content environment.
The logic behind voice and likeness trademarks in the AI context is that copyright law, which typically protects specific creative works, is often an imperfect tool against AI-generated imitation. A synthetic voice that sounds like Taylor Swift is not a copy of a specific Swift recording, so it may not trigger traditional copyright infringement. Trademark law, by contrast, protects identifiers of commercial source and can be applied when someone uses a mark in commerce in a way likely to cause consumer confusion.
Registering a celebrity's own voice phrase or visual identity as a trademark gives rights holders a statutory basis to pursue platforms and developers who commercialize AI-generated content that mimics those marks without authorization.
Why Now | AI Deepfakes and the Taylor Swift Precedent
Swift has already been at the center of one of the highest-profile AI deepfake incidents in recent memory. In early 2024, explicit AI-generated images of the artist circulated widely on X (formerly Twitter), prompting platform-level takedowns, Congressional calls for federal legislation, and a public statement from the White House. The episode accelerated legislative discussions around the NO FAKES Act and similar bills aimed at creating a federal right of publicity specifically applicable to AI-generated content.
The trademark filings represent a parallel private-law strategy, one that does not require waiting for Congress to act. Trademark registration provides a standing basis for federal litigation and puts platforms on notice that use of Swift's trademarked identity markers in AI-generated content may constitute infringement.
TAS Rights Management has been aggressive about protecting Swift's intellectual property on multiple fronts. The entity previously filed trademarks on album titles, tour names, and merchandise categories, establishing a broad IP portfolio around the Swift commercial brand.
Legal Context | What Trademark Covers and What It Doesn't
Trademark protections for voice and likeness are not unlimited. The filings cover commercial use of those marks in trade, meaning the strongest protections apply when someone uses a Swift voice clone or image to sell a product, generate revenue from advertising, or trade on consumer association with her brand.
Parody, commentary, and non-commercial uses retain First Amendment protection. A fan account that posts an AI voice clip for entertainment purposes occupies a different legal position than an advertiser who uses a Swift voice clone to sell a product without authorization.
The image trademark is narrower than a general likeness right. It covers the specific described photograph, not every possible visual representation of Swift. A right of publicity claim, which varies by state under current U.S. law, would be the more applicable tool for broader likeness protection. Texas, California, and New York all have right of publicity statutes with varying scopes. A federal NO FAKES Act, if enacted, would standardize that protection nationally.
Industry Watch | Who Else Is Filing
Swift and McConaughey are part of a broader wave. IP attorneys tracking the USPTO have noted increased filings from celebrities and their management entities over the past 18 months covering voice, signature phrases, and visual marks as AI voice cloning tools have become commercially accessible. Musicians, actors, and athletes have all entered the filing trend, anticipating federal legislation that may require existing trademark or publicity rights as a threshold for seeking relief.
For the full landscape of IP and copyright litigation in entertainment and tech, see the copyright and IP hub. Related coverage: Baltimore's lawsuit against xAI over Grok-generated deepfakes and the Ninth Circuit App Store ruling pause.
Entertainment and IP reporter at ObjectWire. Covers copyright, trademark, and AI law in music, gaming, and film.