District court dismisses authors’ claims for direct copyright infringement based on derivative work theory, vicarious copyright infringement and violation of Digital Millennium Copyright Act and other claims based on allegations that plaintiffs’ books were used in training of Meta’s artificial intelligence product, LLaMA.
Meta developed and released, on a limited basis, its artificial intelligence software Large Language Model Meta AI (LLaMA). Meta trained LLaMA by exposing it to massive amounts of text from various sources. Among those sources were books authored by Sarah Silverman, Richard Kadrey and Christopher Golden. Silverman, Kadrey and Golden brought a putative class action against Meta alleging six causes of action: direct copyright infringement, vicarious copyright infringement, violation of the Digital Millennium Copyright Act (DMCA), unfair competition, unjust enrichment and negligence. On Meta’s motion, the court dismissed all the claims with the exception of plaintiffs’ assertion that Meta’s alleged unauthorized copying of the plaintiffs’ books to train LLaMA constituted direct copyright infringement, which defendant did not move to dismiss. The dismissal of all the claims save the negligence claim was granted with leave to amend; the negligence claim was dismissed with prejudice.
Meta moved to dismiss plaintiffs’ cause of action for direct copyright infringement only to the extent that it was premised on a theory that the software comprising LLaMA is itself an infringing derivative work. The court found this theory of liability to be “nonsensical,” as the models themselves were in no way “a recasting or adaptation of the plaintiffs’ books.” Plaintiffs’ claim for vicarious copyright infringement failed because the complaint did not allege that any output generated by LLaMA contained protectable expression that recast, transformed or adapted the books. Without “an infringing output, there can be no vicarious infringement.” Citing the ruling in Andersen v. Stability AI Ltd. (see our summary here), the court held that allegations of substantial similarity are required in order to support a theory of derivative infringement; the plaintiffs would need to allege that the outputs of the software “incorporate in some form a portion of” their books. And, in alignment with the court in Andersen v. Stability AI Ltd., the court here found the rationale and holding in Range Road Music, Inc. v. East Coast Foods, Inc. to be unavailing in circumventing the requirement to adequately allege (and ultimately prove) substantial similarity. Plaintiffs’ claims under the DMCA failed because they did not allege that LLaMA generated and distributed copies of the books. Meta therefore could not have distributed the books with false copyright management information, as alleged. Moreover, the software comprising LLaMA itself was not an infringing derivative work.
The court concluded that the Copyright Act preempted each of plaintiffs’ state-law claims for unfair competition, unjust enrichment and negligence. Each of these causes of action relied on the same rights contained in the Copyright Act. Plaintiffs’ sole remaining cause of action is for direct infringement based on Meta’s allegedly unauthorized copying of plaintiffs’ books when training LLaMA.
Summary prepared by Melanie Howard and Keane Barger.