IP Law Daily, COPYRIGHT—N.D. Cal.: Alphabet is out of the soup in Google AI infringement case, (Sep 15, 2025)

By Matthew Hersh, J.D.

The complaint alleged nothing more than a traditional parent-subsidiary relationship between the two companies.

The named plaintiffs in a putative class action over Google’s generative artificial intelligence services could not rope Google’s parent company Alphabet into the proceedings, the federal court for San Jose, California has held. The court, in granting a motion to dismiss the parent company from the action, underscored that the named plaintiffs identified nothing that would tie Alphabet to the allegedly infringing conduct beyond the mere fact that Alphabet and Google had a traditional parent-subsidiary relationship (In re Google Generative AI Copyright Litigation, No. 5:23-cv-03440-EKL (N.D. Cal. Sept. 11, 2025)).

The court’s most recent action comes in a consolidated action that originated as two separate putative class actions against Google and its parent company Alphabet Inc. The ten named plaintiffs in the action allege that Google used their works to train a series of popular generative artificial intelligence models, including the well-known Gemini and Imagen models. They also claim that Google is working to incorporate those AI models into a wide range of other Google products. Their consolidated complaint proposed a class of all U.S. residents with valid copyright registrations whose works were infringed by Google to train its AI model. (In April of this year, the district court found that the plaintiffs could move forward with their class action, but only if the class definition were narrowed to the more objection definition of U.S. residents whose copyrighted works were “used by Google to train Google’s Generative AI Models” during the relevant class period).

Google and Alphabet moved to dismiss the claim, leading to this opinion.

Copyright registration allegations. The court rejected Google’s attempt to dismiss all of the direct allegations brought by one of the named plaintiffs, Sarah Anderson, on the basis of lack of registration. Anderson alleged that all of the works at issue in the lawsuit were properly registered, but she made those registrations in compilation form—and her registration certificates expressly carved out certain previously published works from the compilation registration. Because it was not clear from the registration certificates whether the works in suit were or were not included in those carveouts, Google argued, the allegations of registration were inadequate. But “[t]he Court has no basis to infer,” the court held, “that because some materials in the compilation were previously published, the works asserted here were previously published.” The claim would survive the motion to dismiss.

Nor would the court address, at the current stage, the allegations that some of the identifiable works in suit were carved out from the registration because they had been posted to X prior to the compilation registration. “This argument would require the Court to take judicial notice of posts on X and X’s terms of use to determine whether Andersen authorized other X users to make copies of her works,” the court noted. “This inquiry is too fact-bound at this stage.” Whatever Anderson posted, and whether the posts constituted publication of the works for copyright registration purposes, the court concluded, would have to be revisited “on a more complete record.”

Allegations of unspecified other works. The court also rejected Google’s claim to dismiss the infringement claims as to “other works” not expressly identified in the complaint. “At the pleading stage, particularly when the scope of the alleged infringement is wide-ranging and not precisely known, it is permissible to plead infringement of representative works,” the court noted. “Here, Plaintiffs trace specific copyrighted works to specific training datasets,” the court found. “These allegations put Google on notice of Plaintiffs’ claims.”

Allegations as to certain models and products. But although the claim would go forward against Google as to most claims and products, certain of Google’s AI products would escape scrutiny in the lawsuit. The named plaintiffs had plausibly alleged copyright infringement as to Google’s PaLM, GLaM, LaMDA, Bard, Gemini, and Imagen models, the court found, because they had identified “specific copyrighted works that were allegedly included in datasets used to train these models.” In contrast, the court noted, the plaintiffs had failed to plausibly allege copyright infringement as to the models known as Codey, Chirp, Veo, MedLM, LearnLM, SecLM, Gemma, CodeGemma, RecurrentGemma, and PaliGemma. Indeed, the only time that the plaintiffs mentioned those models, the court noted, was in “a long list of what appears to be every generative AI model that Google has ever developed.” That would not be sufficient, the court concluded, to keep those models in the complaint.

Vicarious copyright infringement and Alphabet. Although Google did not succeed on its motion to dismiss, its parent company Alphabet would have better luck. The class plaintiffs did not level specific infringement allegations against Alphabet directly. Rather, they argued that Alphabet was vicariously liable, due to its relationship with Google, for its subsidiary’s alleged infringement. But vicarious copyright infringement claims, like other vicarious liability claims, the court noted, “are subject to the bedrock principle of corporate law that a parent company is not liable for a subsidiary’s acts based solely on the parent-subsidiary relationship.” Here, the court noted, the plaintiffs did allege any facts regarding Alphabet’s involvement in the AI training process and offered only “the conclusory allegation that Alphabet had the authority to implement copyright compliance measures.” This allegation, at most, the court found, spoke only to “Alphabet’s alleged right to control Google generally.” That would not suffice to keep Alphabet in the case.

The Case is No. 5:23-cv-03440-EKL.

Judge: Lee, E.

Attorneys: Lesley Elizabeth Weaver (Bleichmar Fonti & Auld LLP) for Jill Leovy. David H. Kramer (Wilson Sonsini Goodrich & Rosati) for Alphabet Inc. and Google LLC.

Companies: Alphabet Inc.; Google LLC

MainStory: TopStory Copyright TechnologyInternet CaliforniaNews GCNNews