California state court upholds radio and podcast host David Greene’s right of publicity, misappropriation, unfair competition and unjust enrichment claims alleging that Google’s use of his voice as the default male voice for its NotebookLM product and in related marketing constituted unlawful use of his “voice” in violation of California’s statutory and common law right of publicity laws, and that the claims are not preempted by the Copyright Act.
Plaintiff David Greene, who hosted NPR’s Morning Edition and the podcast Up First and currently hosts the podcast Left, Right & Center, sued Google LLC and its parent, Alphabet Inc., asserting California state law claims related to Google’s use of recordings of his voice to create the default male voice for the “Audio Overviews” feature of its NotebookLM product, and Google’s use of that voice to market and sell the product, without his permission. Google’s Audio Overviews, introduced in September 2024, turn uploaded documents into a podcast-like conversation between two AI “hosts,” one male and one female. Greene alleged that Google trained its model on his voice or recordings of it and that the male host in NotebookLM renders text into speech “calibrated to replicate Mr. Greene’s unique vocal qualities, cadence, tone, pitch, and expressive style.” In the alternative, he alleged that “regardless of the precise technical means used,” Google deliberately made the male host sound so much like him that ordinary listeners believed the voice was his.
Greene asserted four causes of action: (1) statutory right of publicity under California Civil Code Section 3344, (2) unfair competition under California Business and Professions Code Section 17200, (3) common law right of publicity and (4) unjust enrichment/quasi-contract.
Google removed the case to federal court on a complete preemption theory, but the Northern District of California concluded that the claims fall outside copyright’s subject matter, because a person’s voice and identity are not copyrightable. (Read our summary of the district court’s decision here.)
After the case was remanded to state court, Google demurred to all four claims, arguing that Section 3344 does not apply to an AI-generated voice that merely resembles Greene’s, that the complaint fails to state any claim, that the Copyright Act preempts every claim and that Alphabet is not a proper defendant. The court overruled Google’s demurrer to every cause of action but sustained Alphabet’s demurrer with leave to amend.
The court first turned to whether Greene adequately alleged that the use of his voice stated a claim under Section 3344. The court rejected defendants’ argument based on Midler v. Ford Motor Co., in which the Ninth Circuit held that the statute was “of no aid” to Bette Midler because the sound-alike singer Ford hired used another singer’s voice, not Midler’s. The court distinguished Midler, explaining that Greene’s complaint alleges that Google trained its voice model on recordings of Greene’s own voice, adjusted its output to duplicate his voice and generated a voice that is “unquestionably Mr. Greene’s voice.” According to the court, “the output is not someone else’s voice that happens to resemble Greene’s. It is a digital rendering of Greene’s voice created from Greene’s voice.”
The court found support in the text of Section 3344, which covers the knowing use of another’s “voice … in any manner” and does not limit “voice” to a particular unmodified recording. The court held that this text provided no reason to distinguish a digital reproduction of a voice generated from recordings of that voice from a reproduction of a person’s image, which triggers Section 3344.
The court also rejected defendants’ argument that the California legislature’s treatment of “digital replicas” shows that Section 3344 excludes them. The court took judicial notice of SB 1111, which the California governor signed on Sept. 30, 2026, and which adds a new subdivision (f) to Section 3344, providing that “a voice or likeness includes a digital replica.”
The court further held that Greene adequately pleaded a direct connection between the unauthorized use and Google’s commercial purpose in order to state a claim under Section 3344. It explained that Greene was not required to plead specific facts showing that his identity was a substantial factor in sales. Instead, the claim applies to uses “on or in products” and to uses “for purposes of advertising or selling.” The court held that Greene’s complaint alleges both uses by alleging that his voice is the default male host in a subscription product, that Google chose it “to create familiarity for seasoned podcast listeners and, in effect, maximize Google’s profits,” and that Google used Greene’s voice in demonstrations, promotional materials and a Spotify partnership promoting NotebookLM.
Turning to the common law misappropriation claim, the court noted that the elements are (1) the defendant’s use of the plaintiff’s identity; (2) the appropriation of the plaintiff’s name or likeness to the defendant’s advantage, commercially or otherwise; (3) lack of consent; and (4) resulting injury.
As to the first prong, defendants argued that Greene was required to plead facts showing that his voice is distinctive and widely known, that Google deliberately imitated it to sell a product and that people familiar with his voice believed Greene himself performed the AI output. The court rejected the last of these as a pleading element but held that Greene satisfied the first three. The court held that Greene adequately alleged that his voice was distinctive and well-known, because a radio host “reaches the audience through the voice.” As to deliberate imitation, the court relied on allegations that Google chose Greene’s voice to create familiarity for podcast listeners, built the voice from his recordings, calibrated it to replicate his vocal traits and knew he was recognizable.
The court likewise held that Greene adequately pleaded appropriation of his voice to Google’s advantage. The court rejected Google’s reliance on federal AI training decisions that require plaintiffs to identify the specific works used to train a model. The common law tort does not, the court explained, turn on training but instead requires only deliberate imitation of a distinctive voice to sell a product. The court similarly found the complaint’s allegations that Google uses Greene’s voice as a default host in a paid product, in promotions and in a commercial partnership sufficient to allege Google’s advantage.
The court similarly held that Greene pleaded resulting injury by alleging that Google usurped the market value of his voice and that the false association with NotebookLM has reduced the value of his voice as a licensable asset, impaired his ability to license it to others and produced AI-generated podcasts that compete with his work.
Turning to Greene’s unfair competition claim, the court first held that Greene had standing because the alleged diminution of the commercial value of Greene’s identity, which California treats as a property interest, is a cognizable economic injury. Greene also pleaded unlawful conduct through the “predicate” violation of Section 3344.
The court similarly upheld Greene’s unjust enrichment claim as a quasi-contract claim for restitution. It held that Greene’s allegations of Google’s use of the commercial value of his voice to develop, promote and sell NotebookLM without consent or payment adequately pleaded that Google unjustly retained a benefit.
The court next considered Google’s assertion of copyright preemption as to all four claims. The two-part copyright preemption test asks (1) whether the subject matter of the state law claim falls within the subject matter of the Copyright Act and (2) whether the rights asserted under state law are equivalent to the exclusive rights under 17 U.S.C. Section 106. Both elements must be satisfied for preemption to apply.
The court held that copyright preemption did not apply because Greene’s claims rest upon the alleged appropriation of his voice and identity, neither of which is copyrightable. Agreeing with the federal court that had previously held preemption inapplicable in its remand order, the state court noted that Greene’s claims are “fundamentally about the program’s outputs,” not its inputs. The court reasoned that Greene alleges not that Google reproduced or distributed his recordings but rather that it generated new speech in his voice that he never recorded. Copyright would not reach that conduct, the court noted, “even if Greene owned every recording he made,” since a sound-recording copyright does not extend to independent sounds that imitate the recording. The court acknowledged that any theory treating the copying of Greene’s recordings as independently actionable under state law may be preempted. However, because each cause of action Greene asserted rested on the appropriation of his voice and identity, which were not copyrightable, the court declined to dismiss all claims on preemption grounds.
Finally, the court turned to Alphabet’s argument that the complaint failed to state a basis for holding it responsible as Google’s alter ego. The court agreed, holding that the complaint’s alter ego and agency allegations were conclusory. The complaint lacked any facts showing that honoring Google’s separate existence would promote injustice or that Alphabet participated in developing, training, deploying or promoting the male host’s voice, the court explained. The court accordingly dismissed all claims as to Alphabet with leave to amend.
Summary prepared by Tal Dickstein and David Forrest
Download our Intellectual Property/Entertainment Cases of Interest mobile app using the links below. 
Greene v. Google LLC – Loeb & Loeb LLP
Leave a Comment
