Voice and Sound-Alikes

Your voice is not copyrightable. That single fact controls the entire subject. Copyright protects original works of authorship fixed in a tangible medium of expression under 17 U.S.C. § 102(a), and a voice is neither: it is not a work, and the speaker does not fix it. Protection against sound-alikes therefore comes from the right of publicity, a state law tort against commercial appropriation of identity.

That relocation changes everything downstream. You do not register a voice, there is no term of years, there is no federal filing, and the analysis does not ask what was copied. It asks whether someone took your identity and sold something with it.

Three provisions close the door in sequence.

Section 102(a) requires fixation in a tangible medium, and 17 U.S.C. § 101 defines a work as “fixed” only when its embodiment in a copy or phonorecord happens “by or under the authority of the author.” When you speak, nothing is fixed. When a studio records you, a sound recording exists, but the copyright belongs to whoever authored and fixed that recording, and it covers that recording, not the vocal identity audible in it.

Section 102(b) excludes systems, methods, and concepts. The timbre, rasp, phrasing, and attack that make a voice recognizable are attributes, not expression.

Most directly, 17 U.S.C. § 114(b) states that the rights in a sound recording “do not extend to the making or duplication of another sound recording that consists entirely of an independent fixation of other sounds, even though such sounds imitate or simulate those in the copyrighted sound recording.” Congress wrote the sound-alike exemption into the statute. Hire a session singer, record fresh audio that sounds exactly like the original, and you have not infringed the record. That is not a loophole. It is the design.

The Ninth Circuit made the consequence explicit in Midler v. Ford Motor Co., 849 F.2d 460 (9th Cir. 1988): “A voice is not copyrightable. The sounds are not ‘fixed.’”

The rule from Midler v. Ford

Ford’s agency wanted Bette Midler’s 1973 recording of “Do You Want to Dance” for a Mercury Sable campaign. Midler declined. The agency licensed the composition from the copyright owner, then hired one of Midler’s former backup singers and instructed her to sound as much as possible like the Midler record. Listeners believed it was Midler. It was not.

California Civil Code § 3344 protects name, voice, signature, photograph, and likeness, but the court held it did not reach this conduct because the defendants used an imitator’s actual voice, not Midler’s. The common law claim survived anyway. The holding, quoted constantly since:

“When a distinctive voice of a professional singer is widely known and is deliberately imitated in order to sell a product, the sellers have appropriated what is not theirs and have committed a tort in California.”

Every clause is load-bearing. Distinctive. Widely known. Professional singer. Deliberately imitated. To sell a product. On remand, a jury awarded Midler $400,000 against the agency alone, affirmed in Midler v. Young & Rubicam, Inc., 944 F.2d 909 (9th Cir. 1991). Ford was out of the case by the time of trial; the agency, which chose and coached the sound-alike, carried the verdict.

Waits v. Frito-Lay and what deliberateness costs

Waits v. Frito-Lay, Inc., 978 F.2d 1093 (9th Cir. 1992), took the rule from principle to price. Tom Waits, who publicly refused to do commercials on principle, was imitated in a SalsaRio Doritos radio spot by a singer hired specifically because he could do Waits. The agency had internally flagged the risk and proceeded.

The jury awarded $375,000 compensatory damages, broken into $100,000 for the fair market value of his services, $200,000 for injury to his peace, happiness, and feelings, and $75,000 for injury to goodwill and future publicity value, plus $2 million in punitive damages for voice misappropriation and $100,000 on a Lanham Act false endorsement claim. The Ninth Circuit affirmed the voice misappropriation verdict in full, including the punitive award, and vacated only the Lanham Act damages as duplicative of the fair market value component.

Waits did two things. It confirmed that Section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a), supports a false endorsement claim for a distinctive imitated voice, giving plaintiffs a federal theory alongside the state tort. And it established that a knowing sound-alike is the kind of conduct juries punish. The punitive award was more than five times compensatory.

The limits nobody mentions

Sound-alikes are not unlawful as such. Impressionists work legally. Tribute bands exist. The doctrine has real edges:

  • Distinctiveness is required. A generic voice, however well imitated, supports nothing. Courts look for a voice the public identifies with one person.
  • Wide recognition is required. Midler speaks of a voice “widely known.” An unknown performer with an unusual voice has a much harder claim.
  • Commercial use is the trigger. The tort targets selling goods and services with someone’s identity, not depicting, discussing, parodying, or performing.
  • Copyright preemption cuts real claims. Laws v. Sony Music Entertainment, Inc., 448 F.3d 1134 (9th Cir. 2006), preempted a singer’s publicity claim where the defendant used her actual licensed recording in a sample. When the complaint is about a recording, copyright occupies the field. Midler survives only because a voice is outside copyright’s subject matter.

Sinatra v. Goodyear Tire & Rubber Co., 435 F.2d 711 (9th Cir. 1970), is the cautionary case. Nancy Sinatra sued over a tire ad using “These Boots Are Made for Walkin’” performed by other singers. She lost. Goodyear had licensed the composition, and as the court read her complaint, her grievance was not that “her sound is uniquely personal” but that the song plus arrangement identified her. That theory would let any performer sue any licensee, which copyright cannot tolerate. Frame the claim around the work and you lose. Frame it around the voice and you are in Midler.

Where AI voice cloning lands

Cloning collapses the old economics: what once required a skilled impersonator now takes seconds of sample audio. The doctrine, though, translates cleanly, because it never depended on how the imitation was produced.

Lehrman v. Lovo, Inc. (S.D.N.Y. 2025) is the first substantial test. Two voice actors were hired through Fiverr by people who turned out to be Lovo employees, told the recordings were for internal research, and later found their voices sold as AI clones. The court dismissed the Lanham Act false association claim, reasoning that their voices were the product rather than a source identifier, dismissed false advertising because Lovo truthfully marketed the voices as clones, and dismissed most copyright claims. It allowed the New York Civil Rights Law §§ 50 and 51 publicity claims and breach of contract claims to proceed, holding that New York’s digital replica provision could plausibly reach an audio-only clone.

The pattern repeats the 1988 result: trademark and copyright fail, publicity survives. Legislatures have moved to widen that surviving lane. Tennessee’s ELVIS Act, effective July 1, 2024, added voice expressly to its Personal Rights Protection Act and reaches tools whose primary purpose is producing unauthorized replicas. California’s AB 2602 and AB 1836 (2024) regulate digital replica contract terms for performers and replicas of deceased personalities. The federal NO FAKES Act remains a proposal, so the map stays state by state.

Frequently asked questions

Can you copyright your voice? No. Copyright attaches to original works fixed in a tangible medium under 17 U.S.C. § 102(a), and a voice is neither a work nor fixed by the speaker. Section 102(b) also excludes systems, methods, and concepts, and the qualities that make a voice recognizable are attributes rather than expression. A specific recording of you speaking is copyrightable as a sound recording, but the vocal identity inside it is not. Protection for the voice itself comes from the right of publicity, a state law tort, not from copyright.

Is it illegal to use a sound-alike voice? Not by itself. Imitation is not the wrong. Under Midler v. Ford Motor Co. (9th Cir. 1988), liability arises when a professional singer’s distinctive and widely known voice is deliberately imitated to sell a product. Ordinary impressions, unknown voices, non-commercial imitation, and parody generally fall outside the rule. The claim is commercial appropriation of identity, not mimicry.

Why did Bette Midler win but Nancy Sinatra lose? Sinatra v. Goodyear (9th Cir. 1970) failed because Goodyear had licensed the composition and Sinatra’s theory depended on the song rather than on her voice as such, which put her claim in conflict with copyright. Midler complained about her voice itself being imitated after she declined the job, an interest copyright does not reach. The distinction is voice as identity versus voice as attached to a licensed work.

Does licensing a song protect you from a voice claim? No. Licenses to the musical composition and the sound recording come from different owners and cover different rights. Neither conveys any right to imitate a performer’s voice. Ford’s agency licensed “Do You Want to Dance” from the copyright owner and still committed the tort in Midler v. Ford Motor Co. (9th Cir. 1988), because the wrong was the deliberate imitation of the voice, not the use of the song. Sinatra shows the mirror image: build the claim on the licensed work instead of the voice and copyright defeats it.

Authorities and sources

Going further: AI Voice Cloning and Deepfakes, what your rights actually are .

This page is general legal information, not legal advice, and it does not create an attorney-client relationship.

The cases behind this
Voice & Sound-Alikes

Hugga-Hugga, Brrr, and the Limits of Voice: Tin Pan Apple v. Miller Brewing

When a beer ad hired Fat Boys imitators, a New York court let the look-alike, copyright, and Lanham Act claims proceed but held that copying a distinctive voice did not, by itself, violate the state privacy statute as it then read. The legislature added voice in 1995.

June 23, 2026
Practical Guides
More in Right of Publicity