In January, we covered Matthew McConaughey’s strategy of registering his iconic catchphrase, “Alright, alright, alright,” as a sound mark to fight AI misuse. That strategy just got a major sequel. On April 24, 2026, Taylor Swift’s company filed three trademark applications with the USPTO: two sound marks covering audio clips of her voice saying, “Hey, it’s Taylor” and “Hey, it’s Taylor Swift,” and one image mark featuring a photograph of her performing on stage. The filings represent a broadening of the celebrity trademark playbook, signaling that prominent figures in entertainment see trademark law as a frontline weapon against AI-generated deepfakes.

Why Trademark Law? Filling the Gaps

Copyright law protects original recordings but does not necessarily reach AI-generated audio that mimics a voice without copying a specific recording. State right-of-publicity laws vary significantly in scope and enforceability. Federal trademark registration offers something neither can: nationwide rights, statutory damages, and the ability to invoke customs enforcement against infringing goods at the border. By registering distinctive vocal phrases and images as trademarks, celebrities create a federal enforcement mechanism that augments existing name, image, and likeness (NIL) protections and can be deployed against anyone using an AI-generated replica in commerce.

How It Works: Sound Marks, Image Marks, and the Registration Process

Sound marks require a detailed written description of the audio specimen and a corresponding audio file demonstrating use in commerce. Image marks function similarly to standard trademarks but claim a specific photograph or visual depiction of the applicant. To succeed, the applicant must show the mark is distinctive, identifies the source of goods or services, and is used in commerce. Swift’s filings identify entertainment and merchandising services as the associated goods and services.

Strengths and Limitations

This strategy is most effective as a deterrent and enforcement notice mechanism. The threat of federal litigation, including treble damages under 15 U.S.C. §1117 of the Lanham Act, may discourage bad actors from deploying deepfake content commercially. However, trademarks protect marks as used in connection with specific goods or services, not in all contexts. Purely noncommercial deepfakes (such as parody or political commentary) may fall outside trademark’s reach. Additionally, marks that are narrowly defined to specific audio clips or images may not cover all variations of AI-generated likenesses.

What This Means for You

Athletes, entertainers, public figures, and brands with distinctive voices or visual identities should evaluate whether trademark filings can supplement their existing publicity rights and contractual protections. Identifying your most recognizable vocal phrases, catchphrases, or signature images is the first step. Early registration creates a clean enforcement record and puts potential infringers on constructive notice. As AI-generated content becomes increasingly sophisticated, proactive trademark filings represent one of the most concrete tools available today for protecting personal brand identity. The law is still catching up to the technology, but those who act now will be best positioned as the framework matures.