Technology

Taylor Swift's Bold Trademark Move Against AI Voice Copies

Taylor Swift filed trademark applications to protect her voice and likeness from AI impersonation. The move tests whether trademark law, originally designed for jingles and logos, can stop AI deepfake

Martin HollowayPublished 2w ago4 min readBased on 9 sources
Reading level
Taylor Swift's Bold Trademark Move Against AI Voice Copies

Taylor Swift's Bold Trademark Move Against AI Voice Copies

Taylor Swift filed three trademark applications with the U.S. Patent & Trademark Office on April 24, 2024. Two protect audio clips of her voice saying "Hey, it's Taylor" and "Hey, it's Taylor Swift." The third protects a photograph of her on stage from her Eras Tour, wearing a pink guitar, a multi-colored iridescent bodysuit, and silver boots.

This is one of the highest-profile legal efforts to stop AI-generated impersonations of a celebrity. Swift's move tests new ground — it uses trademark law in a way the law wasn't originally designed to handle.

What Are Sound Marks?

A sound mark is a type of trademark that protects a distinctive sound rather than a name or image. You've heard examples: a jingle in a car ad, or the distinctive chime that plays when your email arrives. These sounds have become so linked to a company or brand that the sound itself can be legally protected.

Swift's voice falls into this category. The trademark applications specifically target protection against AI voice synthesis and deepfakes — technology that can copy someone's voice and make them say things they never said. As AI voice tools have become cheaper and easier to use, the problem has grown quickly.

Why This Matters

Swift is not the first celebrity to try this strategy. Matthew McConaughey filed similar trademark applications in January 2024, a few months before Swift. His applications covered images, video, and audio of himself. McConaughey's filings suggest that celebrities and their legal teams believe existing laws may not be strong enough to stop AI impersonations.

The U.S. Patent & Trademark Office has noticed. In August 2024, the office held a meeting with industry experts to discuss "Name, Image, and Likeness Protection in the Age of AI." The meeting shows that government agencies are thinking about how old legal tools should work in a world where AI can copy a person's voice or likeness convincingly.

The Legal Problems Ahead

There are real obstacles to Swift's strategy. Trademark law traditionally requires that a sound function as a "source identifier" — meaning the public recognizes it and knows it comes from a particular company or person. A jingle on a Coca-Cola ad clearly does this. But a person's speaking voice appears in so many contexts — interviews, podcasts, home videos, casual conversations — that proving it works as a trademark identifier is much harder.

A bigger question looms: does an AI-generated voice count as trademark infringement at all? Trademark law was built to stop companies from confusing customers about who made a product. Many AI voice syntheses happen online, in parodies, or in personal videos — places where trademark law traditionally does not reach. Courts may decide these uses fall outside what trademark law was designed to protect.

The Problem With Enforcement

Even if the trademark applications succeed, catching infringements will be difficult. When a store illegally sells a counterfeit handbag, trademark holders know where to look. But AI-generated content comes from countless apps, websites, and individual users scattered around the world. Finding and stopping it at scale would be far harder than policing a typical trademark violation.

The broader context here is worth noting. We saw similar legal uncertainty decades ago when the internet first arrived. At that time, copyright law — designed for printed books and films — suddenly had to handle digital copying and sharing. Existing laws bent and adapted, but the process took years and many court battles. Swift's voice trademark battle may unfold in much the same way.

What Comes Next

The real test will come when the U.S. Patent & Trademark Office examines these applications. If approved, they may be challenged in court. Those outcomes will help define whether trademark law can evolve to handle AI-generated content, or whether we need brand new laws to protect people's voices and likenesses in the age of artificial media.

Swift is not alone in seeking protection. The music industry has pursued copyright lawsuits and licensing deals. Actors have pushed for stronger protections in union contracts. Swift's approach — using trademark — is different, and it may prove faster or stronger than copyright in some cases. Time will tell whether it works.