Videos

Matthew P. Hintz: So McConaughey, through his entity J.K. Livin Brands Inc., registered a combination of sound marks and motion marks with the USPTO. The sound marks capture the unique pitch and cadence of his delivery of iconic lines “alright, alright, alright" from Dazed and Confused and "just keep livin', right? I mean, what else are we gonna do?"

The motion marks include a seven-second clip of him on a porch and a three-second clip of him in front of a Christmas tree. All eight registrations fall in Classes 9 and 41, covering things like downloadable audiovisual media and entertainment services, including personal appearances by an actor and celebrity.

The core theory is straightforward: if someone uses AI to generate a convincing imitation of McConaughey's voice or likeness, his team can argue that the imitation is likely to confuse consumers into thinking McConaughey endorsed or is the source of that content and bring a federal trademark infringement claim under the Lanham Act. His attorney described this as giving them "a tool now to stop someone in their tracks or take them to federal court."

Bryan Sterba: Why trademarks? Why not rely on copyright or right-of-publicity laws for
this?

Matthew P. Hintz: Right-of-publicity laws are the traditional avenue for this kind of thing—they protect a person's name, image, voice, and likeness from unauthorized commercial use. But they're a patchwork of state laws. Some states have very robust protections, others offer very little, and there's no single federal statute. This creates uncertainty, especially when the allegedly infringing content is distributed nationally on the internet.

Copyright doesn't really help either, because a voice per se is generally not considered an original work of authorship under copyright law. And short catchphrases and phrases typically aren't copyrightable. So McConaughey's team turned to trademarks because federal registration gets you into federal court, gives you a legal presumption of validity and ownership, provides nationwide constructive notice, and opens the door to enhanced damages and injunctive relief under the Lanham Act. It's a stronger enforcement toolkit than what you'd have under most state publicity-rights regimes. His lawyer in an interview put it bluntly: “They don't know what the court will ultimately say, but at least they have a vehicle to test it.”

Bryan Sterba: What's the strongest case for it actually working?

Matthew P. Hintz: The strongest case is this: for someone like McConaughey, certain phrases and mannerisms generally do function as source identifiers in the marketplace. When you hear "alright, alright, alright" in that distinctive drawl, you immediately think “McConaughey”—and by extension, his brand, his entertainment services, his foundation. The phrase serves as an indicia of source in a way that, say, a random person's catchphrase would not.

There's also a dilution argument for famous marks. Under the Lanham Act, the owner of a famous mark can stop uses that blur its distinctiveness or tarnish its reputation, even without a showing of consumer confusion. If McConaughey's marks are deemed "famous," that's a powerful additional avenue—one that doesn't require the traditional likelihood-of-confusion analysis at all. That’s famous for the
mark, not for McConnaughey.

Bryan Sterba: Well alright, alright…I’m just gonna stop there. Why might this strategy ultimately fall short?

Matthew P. Hintz: So, here's the fundamental tension: trademark law is designed to protect indicias of origin. Its purpose is to prevent consumer confusion about the source of goods or services. It's not designed to function as a general purpose right-of-publicity statute or a mechanism for controlling how your
persona is used in the world. 

The problem is that in the most worrisome AI deepfake scenarios—say, someone generates a video of an AI McConaughey delivering a monologue for a YouTube video, or an AI voice clone narrating a podcast—the would-be infringer is typically not trying to confuse consumers into thinking that McConaughey is the source of that video or podcast. The AI-generated McConaughey is the product itself, not a source identifier for someone else's goods or services. And that's a critical distinction. As the court said in Lehrman v. Lovo, "even extremely famous celebrities" cannot use trademark claims "based on the use of their likeness as products rather than as source-identifying marks." 

University of Minnesota law professor [William] McGeveran called this "the wrong tool" to address "a real problem," noting that trademarks "are meant for product-identifiers, not for people." The way an accused infringer uses the mark matters enormously, and trademark law generally cannot reach uses that don't serve a source-identifying function.

Bryan Sterba: So, if I'm an AI researcher and I train my model on McConaughey's voice, or a user generates a McConaughey-sounding output, does that trigger trademark liability?

Matthew P. Hintz: Probably not under current law, and this is another key limitation. Using material protected by trademark registrations as input for an AI model—feeding it training data that includes McConaughey's voice or image—generally is not considered "use in commerce" that identifies the source of
goods or services. Similarly, the AI tool itself isn't "committing" trademark infringement by processing the input and generating output, nor is the developer of the tool likely to be liable for infringement under existing
trademark principles.

Where you might have a case is if someone takes the AI-generated output—a convincing McConaughey voice clone, say—and uses it to promote their own goods or services in a manner that would cause
consumer confusion. For example, if a company used an AI-generated McConaughey voice in an advertisement to make it look like he endorsed their product, that starts to look a lot more like classic trademark infringement—and, frankly, it would probably also be actionable under existing right-of-publicity law without the trademark registration. The registration adds some procedural benefits, but it's not filling a gap that didn't already exist in that scenario.

The gap it's trying to fill—the gray area of semi-creative, semi-commercial AI content monetized through ads on platforms like YouTube or TikTok—is precisely the space where the law is least clear, and where the
trademark's source-identification requirement is hardest to satisfy.

Bryan Sterba: Ok, so to sum up: was this a smart move? And where does this leave
other public figures thinking about AI protection?

Matthew P. Hintz: I think it was a smart tactical move even if it's not a complete solution. The registrations give McConaughey's team stronger cease-and-desist letters, a presumption of validity, and a clearer path into federal court. Even if a court ultimately finds that the trademark doesn't cover a particular use, the existence of the registration creates friction and raises the cost of noncompliance for would-be bad actors. That has real deterrent value. 

But it's important to be honest about the limitations. Trademark protection was designed to protect consumers from confusion in the marketplace, not to give celebrities a veto over any use of their persona. For uses that don't create source confusion—which is likely the majority of AI deepfake
scenarios people are worried about—the trademark may not reach. As the Bloomberg Law analysis put it, McConaughey essentially "stuffed his publicity rights into a trademark box" and is trying to extend its reach
beyond the legal purpose of either. 

The only real answer would be federal legislation. Bills like the NO
FAKES Act would create a federal private right of action against unauthorized AI-generated replicas, which is the direct, purpose-built tool for this problem. Until Congress acts, creative strategies like
McConaughey's are the best practitioners can offer—but they're working around the edges of a legal framework that was not built for this challenge.

Bryan Sterba: Matthew, thank you so much for the valuable insights, and more, importantly tolerating my insufferable puns.

We hope you’ll all join us again next time on "A-I Didn't Know That."