, , , , ,

Judge Lee rejects Anthropic’s motion to dismiss parts of Concord Music’s complaint

No surprise: Judge Lee rejected Anthropic’s motion to dismiss parts of Concord Music’s First Amended Complaint. Judge Lee will decide whether to grant Concord Music’s motion for leave to file a Second Amended Complaint to add DMCA anti-circumvention claim for Anthropic’s alleged “stream ripping” of music tracks from YouTube videos.

1. Contributory infringement: As described above, the allegations provide a plausible basis from which the Court can infer the requisite knowledge for purposes of Publishers’ contributory infringement claim. Therefore, the Court denies the motion to dismiss as to Count II.

    2. Vicarious infringement: Publishers allege that Anthropic “is paid every time . . . end users submit[] a request for Publishers’ song lyrics, and it is paid again every time its Claude API generates output copying and relying on those lyrics.” FAC ¶ 142. Moreover, it is plausible that the availability of Publishers’ lyrics draws customers to use Claude because, as Publishers contend, Claude would not be as popular and valuable as it is but for “the substantial underlying text corpus that includes Publishers’ copyrighted lyrics.” Id. ¶ 146. Indeed, Publishers allege that “[c]ountless users . . . have prompted Claude for [their] lyrics.” Id. ¶ 9, see also id. ¶ 90. These allegations are sufficient at this stage. Considering none of the other vicarious infringement elements are disputed, the Court denies Anthropic’s motion to dismiss as to Count III.

    3. DMCA CMI 1201(b)(1) claim: Anthropic’s cofounders elected to use a tool, Newspaper, that effectively removed CMI from these datasets – and they chose this tool specifically because it was more effective at removing CMI compared to another tool, jusText. Id. ¶¶ 72-73 (asserting that, when Anthropic applied jusText to “a scraped webpage, containing footnotes, a copyright owner name, and ‘© 2019’ copyright notice,” this algorithm did not remove any of this information but, “[i]n contrast, Newspaper, which removed the footnotes, copyright owner name, and copyright notice entirely, was considered ‘a significant improvement.’”). At least some of these cofounders were aware that, during training, large language models tend to memorize “A LOT” of data and then “regurgitate” it. Id. ¶¶ 108-09 & n.29 (quoting Tom B. Brown et al., Language Models are Few-Shot Learners 8-9, ARXIV (July 22, 2020), https://arxiv.org/pdf/2005.14165 [https://perma.cc/25RZ-B8VB]). Taking these allegations together (as true), the Court can plausibly infer that Anthropic had a reasonable basis to know that the dataset curation and training processes it engaged in, whereby it removed CMI (i.e., its intentional removal of CMI), concealed its own infringement.

    4. DMCA CMI 1201(b)(3) claim: Notwithstanding Anthropic’s insistence that the Court follow Tremblay, this case is once again distinguishable. Reply at 11-12. In Tremblay, the court was analyzing whether plaintiffs’ 1202(b)(3) claim was plausibly alleged in connection with third-party user infringing activity. See 716 F. Supp. 3d at 779-780. Whereas, in this case, Publishers allege that Anthropic had a reasonable basis to know that it was concealing its own infringement. See Mango v. BuzzFeed, Inc., 970 F.3d 167, 173 (2d Cir. 2020) (distinguishing situation where “defendant’s own infringement satisfies Section 1202(b)(3)’s second scienter requirement,” from scenario where plaintiff alleges “constructive knowledge of likely future, third-party infringement.”). Likewise, in Tremblay, plaintiffs did “not allege[] that Defendants distributed their books or copies of their books,” nor “that ChatGPT reproduces Plaintiffs copyrighted works without CMI.” 716 F. Supp.3d at 780. That is not the case here, given Publishers allege that Anthropic “distributes
    Publishers’ lyrics and copies of those lyrics knowing that [CMI] has been removed.” See, e.g., id.
    ¶¶ 128, 131. They also allege that “Anthropic wanted to train its Claude AI models specifically on
    the content of Publishers’ lyrics, so that the models’ output would reproduce that expressive
    content,” without CMI. Id. ¶ 74. Thus, Anthropic’s reliance on Tremblay is misplaced and
    Publishers have met their burden under § 1202(b)(3).

    DOWNLOAD JUDGE LEE’S ORDER

    Leave a Reply


    Discover more from Chat GPT Is Eating the World

    Subscribe now to keep reading and get access to the full archive.

    Continue reading