MBW Views is a series of op-eds from eminent music industry people… with something to say. The following MBW op/ed comes from Music Attorneys Heidy Vaquerano (pictured left) and Cynthia Katz (pictured right) at Fox Rothschild. Vaquerano is a partner and co-chair of the Fox Rothschild Entertainment & Sports Law Department in Los Angeles. Katz is Partner and music industry attorney at Fox Rothschild in New York.
There is mounting concern within entertainment industry surrounding the use of artificial intelligence to replicate a celebrity’s voice, their distinctive asset, without consent, compensation, or legal consequence.
But a legal counter-offensive is underway. Rather than relying solely on the patchwork of state right-of-publicity statutes that have traditionally governed the use of celebrities’ names, images and likenesses, a growing list of high-profile entertainers has adopted a new strategy.
They began filing trademark applications with the United States Patent and Trademark Office (“USPTO”), seeking federal protection for the sounds of their voices and iconic catchphrases.
Recent 2026 filings include Taylor Swift, who applied to protect specific spoken-word audio clips tied to her brand, Lionel Richie registered four of his most recognizable lyric lines as sound marks, and the Backstreet Boys filed to cover their group introduction, all following Matthew McConaughey’s trailblazing registrations from 2023.
The argument is that these catchphrases are so iconic and voice-specific that they function like a logo or brand, meriting federal trademark protection because consumers instinctively associate them with a single source.
While creative, the strategy is legally untested and faces an uphill battle: voice has traditionally been protected under a patchwork of state right-of-publicity laws, while trademark law has been reserved for logos, brand names, and other commercial source indicators meant to prevent consumer confusion.
Under the Lanham Act, a mark qualifies for registration if it serves as a “source identifier”: a signal to consumers that a particular good or service originates from a specific source. Traditionally, these are words, logos, and slogans. But the USPTO has long recognized that sounds can serve this function as well. The NBC three-note chime, registered in 1950, is among the oldest purely audio service marks and sound trademarks on the federal register.
“Even if the trademark strategy proves to be a bridge rather than a destination, it forces the legal system to confront who owns a voice in the age of AI.”
The theory advanced by Richie, Swift, and their peers extends this doctrine to a new context: phrases spoken in a specific celebrity’s voice. Matthew McConaughey, whose eight registrations—including his signature “Alright, alright, alright” from Dazed and Confused — were filed beginning in 2023 and granted in late 2025 and early 2026, represents the most advanced example. His portfolio includes not only sound marks but also video-based “motion marks” capturing his distinctive gestures and visual identity.
The legal test for these marks is whether the relevant consuming public associates the sound with a single commercial source. If a listener hears “Alright, alright, alright” in McConaughey’s voice and immediately thinks of McConaughey’s brand, then the sound arguably functions as a trademark.
The trademark approach’s chief attraction is its federal character. The Lanham Act provides a suite of enforcement tools: claims for false endorsement, false designation of origin, and likelihood-of-confusion actions against anyone using a confusingly similar sound in commerce.
If successful, this strategy would be a helpful tool to combat AI because federal registration provides presumptive validity, nationwide jurisdiction, and statutory damages. It is also compelling because it enables platform-level takedowns of AI deepfakes. Google, Amazon, and Meta already remove trademark-infringing listings, and presumably would extend that treatment to registered sound marks — giving celebrities a takedown mechanism without costly litigation.
Trademark law, however, was never designed to protect an entire persona; it protects source identification in commerce by preventing consumer confusion about who is behind a product or service. Stretching it to cover a celebrity’s voice raises unresolved doctrinal questions.
First, a sound mark must signal a commercial source to listeners — not merely be recognizable as a famous voice. Everyone recognizes Matthew McConaughey’s voice, but does hearing it automatically signal a particular commercial origin? The USPTO and reviewing courts will have to resolve that distinction.
Second, the “use in commerce” requirement presents challenges. Trademark owners must demonstrate that the mark has been used as a source identifier in connection with goods or services. A voice used in a song is artistic expression; the same voice used to endorse a product is commercial. The line between these functions may prove difficult to police.
Third, the scope of protection remains uncertain. Even if the USPTO grants these registrations, how much AI replication is sufficient to constitute infringement? Must the AI output use the exact registered phrase, or would any AI-generated content in a voice confusingly similar to the mark holder’s constitute a violation? Courts have historically rejected attempts to stretch trademark protection to cover an entire persona.
The trademark filings of 2025 and 2026 represent something more than a collection of individual legal maneuvers. They signal a broader renegotiation of the relationship between human identity and artificial intelligence.
The right of publicity, the traditional vehicle for protecting celebrity identity, remains a creature of state law, and coverage varies dramatically. Unfortunately, most of these statutes were drafted before AI-generated deepfakes were technologically feasible, leaving significant gaps in coverage.
Legislative reform is also underway:
Tennessee’s ELVIS Act, signed in March 2024, was the first statute to explicitly add “voice” to right-of-publicity protections in the AI context, making unauthorized AI-generated use of a person’s voice a Class A misdemeanor with civil liability. The law includes sensible carve-outs for news reporting, fair use, parody, and criticism.
At the federal level, the NO FAKES Act (Nurture Originals, Foster Art, and Keep Entertainment Safe Act), reintroduced in 2025, passed the Senate Judiciary Committee unanimously by voice vote in June 2026. It would create a new federal intellectual property right giving every individual the right to control their voice and visual likeness in digital replicas, with tiered penalties ranging from $5,000 to $750,000 per violation and a notice-and-takedown framework modeled on the DMCA.
Notably, the bill has garnered support from a coalition including major labels, streaming platforms, and AI companies themselves — Universal Music Group, Sony, Warner, Spotify, Google, OpenAI, and YouTube have all expressed support.
Even if the trademark strategy proves to be a bridge rather than a destination, it forces the legal system to confront who owns a voice in the age of AI. Until stronger protections arrive, celebrities will keep using every tool available, building the legal architecture on which future AI regulation will rest.Music Business Worldwide
