by Angela Dunning, Arminda Bepko, and Jessica Graham
This week saw yet another California federal court dismiss copyright and related claims arising out of the training and output of a generative AI model in Tremblay v. OpenAI, Inc.,[1] a putative class action filed on behalf of a group of authors alleging that OpenAI infringed their copyrighted literary works by using them to train ChatGPT.[2] OpenAI moved to dismiss all claims against it, save the claim for direct copyright infringement, and the court largely sided with OpenAI.
First, the court dismissed plaintiffs’ claim against OpenAI for vicarious copyright infringement based on allegations that the outputs its users generate on ChatGPT are infringing. The court rejected the conclusory assertion that every output of ChatGPT is an infringing derivative work, finding that plaintiffs had failed to allege “what the outputs entail or allege that any particular output is substantially similar – or similar at all – to [plaintiffs’] books.” Absent facts plausibly establishing substantial similarity of protected expression between the works in suit and specific outputs, the complaint failed to allege any direct infringement by users for which OpenAI could be secondarily liable.
The court also dismissed claims for violation of the Digital Millennium Copyright Act (“DMCA”). Plaintiffs failed to allege that OpenAI altered or removed copyright management information (“CMI”), such as author names and copyright notices, from their works with the intent to conceal or induce infringement. Conclusory allegations that “[b]y design, the training process does not preserve any CMI,” do not suffice to state a claim under Section 1202(b)(3) of the DMCA. The court also rejected allegations that OpenAI violated Section 1202(b)(3) – prohibiting distribution of “works” or “copies of works” with the CMI altered or removed – because plaintiffs failed to allege that any of the challenged ChatGPT outputs were identical copies of their works.
Plaintiffs’ claims for negligence, unjust enrichment and violation of the “unlawful” and “fraudulent” prongs of the California unfair competition law (“UCL”) were also dismissed for failure to plead facts plausibly stating a claim. Although the court allowed the claim for violation of the “unfair” prong of the UCL to proceed for now, it noted that the claim may be preempted by the Copyright Act if based on allegations of unauthorized use of plaintiffs’ copyrighted works—setting up a potential basis for dismissal later in the case.
This decision continues the trend of narrowing copyright, DMCA, and related state law claims against generative AI companies. Although the court granted plaintiffs leave to amend, it remains to be seen which of the dismissed claims, if any, plaintiffs will attempt to replead. In other cases challenging the training and output of generative AI models, such as Kadrey v. Meta[3] and Andersen v. Stability AI,[4] courts have dismissed identical claims on largely the same grounds, including lack of alleged substantial similarity, failure to allege removal of CMI from identical works, and copyright preemption. The Kadrey plaintiffs opted not to replead any claim except the core direct copyright infringement claim, and the Andersen plaintiffs abandoned many of these claims in their amended complaint and sought to replead others. Motions to dismiss most claims in the amended complaint in Andersen are currently pending.
Footnotes
[1] Paul Tremblay et v. OpenAI, Inc. et al., No. 3:23-cv-03223 (N.D. Cal. Feb. 12, 2024) (Dkt. 104) (the “Order”).
[2] This case was recently consolidated with a substantively identical case brought by another group of authors, including comedian Sarah Silverman, and the motion addressed both complaints. Silverman et al. v. OpenAI, Inc. et al., No. 3:23-cv-03416 (N.D. Cal. Jul. 7, 2023).
[3] Kadrey et v. Meta Platforms, Inc., No. 23-cv-03417-VC, 2023 WL 8039640 (Nov. 20, 2023) (Order Granting Motion to Dismiss).
[4] Andersen et v. Stability AI Ltd. et al., No. 23-cv-00201-WHO, 2023 WL 7132064 (N.D. Cal. Oct. 30, 2023) (Order on Motions to Dismiss and Strike).
Angela Dunning is a Partner, Arminda Bepko is a Senior Attorney, and Jessica Graham is a Law Clerk at Cleary Gottlieb Steen & Hamilton LLP. The post was first published on the firm’s blog.
The views, opinions and positions expressed within all posts are those of the author(s) alone and do not represent those of the Program on Corporate Compliance and Enforcement (PCCE) or of the New York University School of Law. PCCE makes no representations as to the accuracy, completeness and validity or any statements made on this site and will not be liable any errors, omissions or representations. The copyright of this content belongs to the author(s) and any liability with regards to infringement of intellectual property rights remains with the author(s).