IP Law Daily, COPYRIGHT—N.D. Cal.: Class action against Google over its AI generative model will move forward for now, (Apr 22, 2025)

By Matthew Hersh, J.D.

But Google will still have a chance to challenge the proposed class at the class certification stage.

A group of copyright holders can move forward with their class action lawsuit against Google over its generative AI intelligence model once their class action allegations are redefined to be measurable against objective criteria, the federal court for San Jose, California, has held. But the court, in allowing the class allegations to move forward on the basis of the amended definition, warned that the named plaintiffs would still have to demonstrate that they could overcome challenges to the proposed class—including challenges to commonality and predominance that have proven hard to surmount in past copyright cases—at an eventual class certification stage (In re Google Generative AI Copyright Litigation, No. 5:23-cv-03440-EKL (N.D. Cal. Apr. 21, 2025)).

The court’s action on the motion to strike comes in a consolidated action that originated as two separate putative class actions against Google. The ten named plaintiffs in the action allege that Google used their works to train its popular generative artificial intelligence models Gemini and Imagen. They also claim that Google is working to incorporate those AI models into a wide range of other Google products. Their consolidated complaint proposes a class of all U.S. residents with valid copyright registrations whose works were infringed by Google to train its AI model.

Google moved to strike the class allegations, leading to this opinion.

Fail-safe class. The court granted the motion to strike. The problem with the proposed class, the court noted, was that it covered only persons whose works were actually infringed by Google. That made it a classic case of a “fail-safe class,” the court noted—that is, a class that is “defined in a way that precludes membership unless the liability of the defendant is established.” Here, for example, the court noted, the court could not determine who is a member of the class without deciding the merits of each potential class member’s claim, “including whether the potential class member has a valid copyright registration, whether Google infringed the class member’s work(s), and whether Google has a valid defense based on fair use or license (among other possible defenses).” Because the fail-safe class definition was “defective as a matter of law,” the court concluded, it was “appropriate to strike it at the pleading stage.”

Leave to amend. But while the court granted the motion to strike, it also made clear that the plaintiffs would be able to move forward on the basis of their proposed modification to the class definition. That proposed class definition would include U.S. residents whose copyrighted works were “used by Google to train Google’s Generative AI Models” during the relevant class period. “Redefining the class in this manner is appropriate,” the court noted, “because class membership can be determined according to objective criteria—that is, ownership of a copyrighted work that was used by Google to train its generative AI models.” Thus, the court noted, membership in the class “would not require an upfront determination by the Court that each potential class member will prevail on the merits of an infringement claim.”

Granting the plaintiffs leave to move forward on this basis was appropriate, the court noted, because it would leave intact Google’s ability to challenge the proposed class at the certification stage. Google argued that copyright claims are not suitable for class-action treatment because their individualized nature traditionally present, in the words of caselaw cited by the search giant, “overwhelming problems with commonality and predominance that in themselves bar certification.” But the cases Google cited for this proposition were all class certification decisions, the court noted—and “they do not hold that class certification is impossible in a copyright case.” Rather, the court noted, issues of commonality and predominance would be more appropriately addressed after discovery and further briefing.

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

Judge: Lee, E.

Attorneys: Joseph R. Saveri (Joseph Saveri Law Firm, LLP) and Brian D. Clark (Lockridge Grindal Nauen PLLP) for Jingna Zhang and proposed class. Ryan J. Clarkson (Clarkson Law Firm, P.C.) and Lesley Elizabeth Weaver (Bleichmar Fonti & Auld LLP) for Jill Leovy and proposed class. Michael S. Kwun (Kwun Bhansali Lazarus LLP) for Google LLC and Alphabet Inc.

Companies: Alphabet Inc.; Google LLC

MainStory: TopStory AINews Copyright TechnologyInternet CaliforniaNews GCNNews