Your Voice Is Your Brand. AI Can Clone It in Seconds. Here Is How Trademark Law Protects You.

 

Imagine scrolling through your social media feed and finding a video of yourself endorsing a product you have never used, promoting a service you have never heard of, or saying something that could destroy your professional reputation. Your face. Your voice. Your mannerisms. All convincing. All fake.

A year ago, that scenario felt theoretical. Today, it is happening to celebrities, athletes, business owners, and professionals at an accelerating rate. The technology behind AI-generated deepfakes has advanced to the point where creating a realistic synthetic video or audio clip of someone takes minutes, not days, and costs almost nothing.

The legal system is struggling to keep up. But a handful of high-profile figures have started using a tool that most people would not immediately associate with identity protection: federal trademark law. Their strategy has real implications for anyone whose name, face, or voice is part of their livelihood.

The Deepfake Problem Is Growing Faster Than the Law

The numbers tell the story. One cybersecurity firm estimated that the volume of deepfakes circulating online grew from approximately 500,000 in 2023 to roughly 8 million by 2025. That growth rate, somewhere around 1,500 percent annually, shows no signs of slowing down.

The victims are not limited to Hollywood. Athletes have been deepfaked into fake endorsement videos. Earlier this year, an AI-altered video of NHL player Brady Tkachuk was posted on TikTok, putting inflammatory political words in his mouth that he never said. Business owners have seen their likenesses used to promote products they have no connection to. And for people whose personal brand is their business, the damage from a convincing deepfake can be immediate and severe.

The traditional legal remedy for this kind of identity theft is the right of publicity. Most states have some version of it, either through statute or common law, and it protects individuals from having their name, image, or likeness used commercially without consent. On paper, it sounds like a perfect fit for combating deepfakes.

In practice, there are serious gaps. The right of publicity is a state-by-state patchwork. Protections vary enormously depending on jurisdiction. Some states offer robust statutory frameworks. Others barely recognize the concept. For deepfake content that is generated in one state, hosted on servers in another, and viewed by people in all fifty, the jurisdictional complexity is a significant obstacle.

Copyright law does not solve the problem either. A specific recording of your voice is copyrightable, but AI does not need to copy any particular recording to imitate you. It generates new content that mimics your voice or appearance without duplicating any protected work. Your voice, standing alone, is generally not considered a copyrightable work of authorship.

The core gap: AI can replicate your identity without copying any existing protected work. It generates something new that sounds and looks like you, and the traditional legal tools were not designed for that scenario.

The Trademark Strategy: What McConaughey and Swift Are Doing

In late 2023, actor Matthew McConaughey’s legal team began filing trademark applications with the U.S. Patent and Trademark Office. They were not registering movie titles or merchandise logos. They were registering audio clips of McConaughey’s voice, including his iconic “Alright, alright, alright” line, as well as video clips of his likeness. Those registrations were approved in December 2025.

The strategy is built on a straightforward legal theory. If specific elements of McConaughey’s identity, his voice, his distinctive phrases, his image, are federally registered trademarks, then anyone who uses AI to generate content that sounds or looks like him is potentially infringing those marks under the Lanham Act. Federal trademark infringement claims open the door to nationwide enforcement, injunctive relief, and the possibility of enhanced damages. His attorney described the registrations as giving them a tool to stop someone in their tracks and take them to federal court.

Taylor Swift followed this playbook in April 2026. Her company filed three trademark applications with the USPTO: two sound marks covering her voice saying “Hey, it’s Taylor Swift” and “Hey, it’s Taylor,” and one visual mark covering a specific image of her performing with a pink guitar. Jimmy Kimmel filed similar applications around the same time.

Why Trademarks Instead of Other Legal Tools?

The appeal of trademark law in this context comes down to several practical advantages that other legal frameworks do not offer.

Federal registration provides nationwide protection in a single filing, eliminating the need to navigate a patchwork of state right-of-publicity laws. It creates a legal presumption of ownership and validity, which shifts the burden in litigation. It provides constructive notice to the world that these identity elements are claimed as source identifiers. And it opens the door to the Lanham Act’s false endorsement provisions under Section 43(a), which prohibit the use of any name, symbol, or device likely to confuse consumers about who is endorsing a product or service.

There are practical enforcement benefits too. Social media platforms and hosting companies tend to respond more quickly to trademark infringement notices backed by federal registrations than they do to state-law publicity claims. When a deepfake is going viral, the speed of takedown can be the difference between manageable damage and a full-blown crisis.

The USPTO has signaled its support for this evolving use of trademark law. The office recently launched a dedicated NIL (Name, Image, and Likeness) resource page designed to educate athletes, entertainers, and entrepreneurs about how federal trademark registration can protect their personal brand assets. That is a meaningful institutional endorsement of the concept.

The Honest Limitations

Trademark protection requires that the registered element function as a source identifier, meaning that consumers associate it with a specific person as the source of goods or services. A common phrase or a generic photo may not meet that threshold. The more distinctive and recognizable the element, the stronger the claim. That is why McConaughey registered a phrase so iconic that virtually everyone associates it with him, and why Swift chose very specific voice clips rather than something general.

This strategy has also not yet been tested in court. Legal scholars have noted that using trademark law to fight AI deepfakes remains untested in litigation. The registrations exist. The legal theory is sound. But no judge has ruled on the question. We are operating on strong reasoning and good precedent from analogous cases, like Waits v. Frito-Lay, where the Ninth Circuit held that unauthorized imitation of Tom Waits’ voice in advertising constituted false endorsement under the Lanham Act. But a direct ruling on AI deepfakes and trademark law has not happened yet.

And trademark claims require a showing of likely consumer confusion. If a deepfake is clearly labeled as parody, commentary, or satire, the First Amendment may shield it regardless of any trademark registration. Trademark law prevents marketplace confusion. It is not a blanket prohibition on all uses of someone’s identity.

Trademark law is a powerful supplement to right-of-publicity claims, not a wholesale replacement. The strongest protection strategy uses multiple legal tools together: trademark registration, right-of-publicity claims, copyright where applicable, and contract provisions that restrict AI use.

What Small Business Owners and Professionals Should Do Now

You do not need to be Taylor Swift to benefit from this kind of proactive brand protection. If your name, face, or voice is part of how you do business, AI-generated deepfakes are a risk you should be thinking about. Here is a practical framework.

Audit your identity assets. Make a list of every name, nickname, tagline, logo, and distinctive element that your clients and audience associate with you. Identify which of those elements functions as a source identifier for your goods or services. Those are the candidates for federal trademark registration.

Register your trademarks. If you have not registered the core elements of your brand identity with the USPTO, now is the time. A federal registration costs relatively little compared to the protection it provides. Focus on the elements that are most commercially valuable and most likely to be targeted by AI misuse.

Add AI clauses to your contracts. Whether you are entering into endorsement deals, content partnerships, licensing agreements, or vendor relationships, include specific language that addresses AI-generated content. Define what is permissible. Require express written consent for any synthetic reproduction of your voice, image, or likeness. These clauses are becoming standard in entertainment contracts, and there is no reason they should not be standard in commercial agreements across every industry.

Document your authentic content. Maintain a record of your genuine photos, video appearances, and audio recordings. If you ever need to demonstrate that a deepfake is fake, having a clear archive of your real content makes the evidentiary case straightforward.

Act fast if you find a deepfake. If you discover AI-generated content using your identity, preserve everything immediately: screenshots, URLs, timestamps, account information, engagement data. Deepfake content can disappear quickly once the creator realizes they have been identified. Preserve the evidence first, then pursue your legal options.

The Bigger Picture

We are living through a period where the technology to replicate a human identity is advancing much faster than the legal frameworks designed to protect it. Right-of-publicity laws were written before AI could clone a voice in seconds. Copyright was designed to protect creative works, not the people behind them. Trademark law was built to prevent consumer confusion in the marketplace, not to serve as a personal identity shield.

But law has always adapted to new technology, even when the adaptation takes longer than we would like. The McConaughey and Swift trademark filings are an example of creative lawyering at its best: taking existing legal architecture and applying it to a threat that the statute’s drafters never imagined. Congress is moving too. The TAKE IT DOWN Act, signed in May 2025, created new federal criminal liability for publishing non-consensual intimate deepfakes. Tennessee’s ELVIS Act directly addresses AI voice replication. And there are active conversations in Congress about creating a comprehensive federal right of publicity that would provide consistent nationwide protection.

In the meantime, the smartest approach is to layer your protections. Trademark registration. Right-of-publicity claims where available. Contract provisions that anticipate AI misuse. And vigilant monitoring of how your identity is being used online.

If you have built a business around your name, your expertise, and your reputation, that brand is one of your most valuable assets. Protecting it from AI exploitation is no longer a hypothetical exercise. It is a business necessity.

Jessica Eaves Mathews is an award-winning trademark, copyright and business attorney, AI governance expert, and founder of Leverage Legal Group, a trademark, copyright, AI, and digital regulatory compliance law firm focused on protecting founders, startups, small and mid-sized businesses, and independent creators. She writes about trademark law, copyright, AI, and brand protection at leveragelegalgroup.com and on LinkedIn.  Follow her on LinkedIn, on Instagram @TopTrademarkLawyer, or on Substack at jessicaeavesmathews.substack.com.