Trade mark as a tool for protection against AI: legal prospects and risks
- Registration details and the specifics of the subject matter
- Key legal challenges
- The Lehrman v. Lovo precedent
- Alternative protection methods
- Conclusion
- Summary and recommendations
In the media space, Matthew McConaughey’s strategy has been actively discussed: he is using trade mark registrations to protect his image, voice and the famous phrase “alright, alright, alright” from use by artificial intelligence. This is presented as an innovative way to combat deepfakes in circumstances where traditional copyright does not guarantee adequate protection.
Despite the media attention, using trade marks to protect celebrities’ persona is not a legal novelty (Paris Hilton and Taylor Swift previously took similar steps). The distinctive feature of McConaughey’s case is merely that “protection against AI” is expressly stated as the rationale, which makes the step more of a strategic PR move than a fundamental change in enforcement practice.
Registration details and the specifics of the subject matter
In December 2023, J.K. Livin Brands Inc. (representing McConaughey’s interests) filed eight trade mark applications, which were approved by the USPTO in December 2025. The portfolio includes:
- A sound mark: the phrase “alright, alright, alright” with a detailed frequency/pitch description (the first syllable of the first two words at a low pitch, the second syllable at a higher pitch, and the first syllable of the last word at the highest pitch). This is the first registration of this type in USPTO practice.
- Video clips: a 7-second clip on a porch and a 3-second clip by a Christmas tree.
- An audio recording: the phrase “Just keep livin’, right?”.
Key legal challenges
Registering a face and voice as a trade mark (TM) does not guarantee automatic protection of a person’s image for the following reasons:
- Scope of use: a TM protects the subject matter exclusively in the course of trade as a means of individualising goods or services. It does not prevent any use of the image outside commercial activity.
- Likelihood of confusion criterion: to establish infringement, it is necessary to prove that use of an AI clone misleads consumers as to the source of the goods. If AI generates a voice similar to the actor’s voice but does not present it as the original and does not use it in advertising, proving deception is extremely difficult.
- Lack of a workable enforcement mechanism: there is a risk that a TM covering an entire person would allow rights holders to unreasonably restrict the activities of other people who have a similar appearance or voice.
The Lehrman v. Lovo precedent
In July 2025, the U.S. District Court for the Southern District of New York, in Lehrman v. Lovo Inc., cast doubt on the effectiveness of this strategy. Voice actors attempted to protect their voices from AI cloning by an AI company through trade mark mechanisms; however, the court dismissed the claims.
The court’s conclusion: celebrities cannot use TMs where their voice or face is the product itself (content), rather than an identifier of the producer of goods. The court drew an analogy with the use of photographs of Babe Ruth in calendars: the image is part of the product’s content, not its trade mark.
Alternative protection methods
Against the backdrop of imperfections in TM mechanisms, sector-specific legislative initiatives are developing:
- ELVIS Act (Tennessee, 2024): the first law defining a voice as property and providing for civil and criminal liability for its unauthorised simulation.
- No AI FRAUD Act (federal level): a bill aimed at creating a nationwide “right of publicity” to protect against AI cloning.
Conclusion
Protection against AI through trade mark registration does not provide significant advantages over copyright or specialised right-of-publicity laws. Even McConaughey’s lawyers (Jonathan Pollack and Kevin Yorn) acknowledge that their position is “not ideal” and the outcome of potential litigation is unpredictable.
This strategy serves not so much as a legal shield against deepfakes as an additional lever in commercial negotiations. Having a registered TM strengthens the actor’s position when entering into contracts and makes cease-and-desist letters more persuasive, but it does not revolutionise the legal landscape.
Summary and recommendations
Registering elements of a persona as TMs is advisable primarily to strengthen negotiating positions in commercial deals. For comprehensive protection against deepfakes and unauthorised cloning, it is necessary to use a suite of tools: from right-of-publicity rules to specialised AI legislation.
Authors: Kamal Tserakhau, Aleksei Molchanov.
Need advice on protecting intellectual property from neural networks?
Contact REVERA experts to audit your assets and develop an individual protection strategy in the digital environment.
