IP Law Daily, PUBLICITY RIGHTS NEWS: Voice actors, authors sue AI audio platform ElevenLabs, (Sep 4, 2024)

By Cheryl Beise, J.D.

ElevenLabs is profiting from cloning the plaintiffs’ voices and training its AI text-to-speech software using copyrighted audio book narrations, the complaint alleges.

Two voice actors, two science fiction authors, and a publisher have filed suit against ElevenLabs, Inc, accusing the software company of violating the voice actors’ publicity rights by using their voices to train their artificial intelligence text-to-speech software. The complaint also accuses ElevenLabs of violating the anticircumvention provisions of the Digital Millennium Copyright Act, or DMCA (Vacker v. ElevenLabs, Inc., No. 1:24-cv-00987-UNA (D. Del. Aug. 29, 2024)).

The plaintiffs in this case are voice actors Karissa Vacker and Mark Boyett, authors Brian Larson and Vaughn Heppner, and publisher Iron Tower Press, Inc., owned by Larson. Vacker, a Texas citizen, and Boyett a New York citizen, provide narrations for audiobooks, advertisers, and film producers. Larson, a California resident, is a best-selling science fiction author of more than 70 books that have sold millions of copies. Heppner, a resident of Nevada, is an author specializing in thrillers and science fiction. Boyett has narrated over 40 of Larson’s books and 20 of Larson’s books.

The defendant, New-York based ElevenLabs, Inc., is a software company that specializes in developing natural-sounding speech synthesis and text-to-speech software using AI and deep learning. It offers “Speech Synthesis” AI-assisted text-to-speech software that allows users to create synthetic audio narrations of written text through its website, elevenlabs.io, and through its API, according to the complaint. The system also includes the ability to “clone” the voice of a user from a sample provided by the user.

In January 2023, ElevenLabs publicly released its beta platform, which offered nine default voices, including “Adam” and “Bella.” By January 2024, ElevenLabs had attracted millions of registered users and had raised over $100 million in funding. Its current V-2 model supports text-to-speech in 28 languages, “emulating the voice, prosody, pacing, and style, in any of the supported languages as if the person was perfectly bilingual,” according to the complaint. ElevenLabs’ Dubbing Studio product enables users to dub entire movies.

ElevenLabs’ system employs generative artificial intelligence to clone voices, synthesize speech, and perform multilingual voice-to-voice translation, according to the complaint.

The complaint asserts that ElevenLabs used recordings of audiobook narrations performed by Boyett and Vacker in the training data used to train its generative AI systems, and trained and/or selected the “Adam” and “Bella” default voices based on audiobook narrations performed by Boyett and Vacker.

ElevenLabs has promoted its services and attracted millions of users by offering professional-quality audio narrations using the voices of Vacker and Boyett without their permission, according to the complaint. The complaint alleges that Vacker is the source of the “Bella” voice, which imitates Vacker’s voice, capturing her unique vocal qualities, intonation, and style. ElevenLabs removed the Bella voice from its platform, but new recordings are available on YouTube, social media, and elsewhere, according to the complaint.

Vacker asserts claims for invasion of privacy, through the misappropriation of her likeness and right of publicity, under Texas common law, and unjust enrichment under Texas law. Boyer asserts claims for misappropriation of his likeness and publicity rights Under New York Civil Rights Law § 51. The authors and publisher assert claims for violation of the DMCA’s anticircumvention provisions, 17 U.S.C. §§1201-1203.

The plaintiffs are requesting a permanent injunction enjoining ElevenLabs from creating and distributing to its users any synthetic audio recordings of audio narrations generated in the likeness of Vacker’s or Boyett’s voices and requiring ElevenLabs to instruct its users not to create further recordings using “Bella,” “Adam,” or similar voice clones; (2) an award of ElevenLabs’ ill-gotten gains, voice-actor plaintiffs’ lost profits, or both—and no less than a reasonable royalty together with interest and costs; (3) an award to the voice-actor plaintiffs for exemplary damages under N.Y. Civ. Rights Law § 51 and Texas law; and (4) an award to the author/publisher plaintiffs for damages and other monetary relief, as a result of ElevenLabs’ violation of the DMCA; (5) reasonable attorney fees; and (6) pre- and post-judgment interest.

The Case is No. 1:24-cv-00987-UNA.

Judge: Andrews, R.

Attorneys: Brian E. Farnan (Farnan LLP) for Karissa Vacker.

Companies: ElevenLabs, Inc.; Iron Tower Press, Inc.

News: AINews Copyright PublicityRights TechnologyInternet DelawareNews