The Kadrey plaintiffs are objecting to Meta’s proposed redactions in discovery related to information about the methods it used to download datasets and the specific datasets used to train its AI model Llama.
Plaintiffs respectfully submit this opposition to Defendant Meta Platforms, Inc.’s proposed redactions to the Court’s December 20, 2024 Discovery Order (Dkt. No. 364) (“Proposed Redactions”). Meta’s heavy-handed redactions to a discovery order do nothing to protect Meta’s commercially sensitive business information, but instead merely continue the troubling pattern of redacting all references to the methods Meta used to download and distribute pirated works from third-party datasets, as well as all references to the specific datasets Meta downloaded. Plaintiffs have already confronted this recurring over-redaction in opposition to a prior Administrative Motion to File Under Seal, which is pending. See Dkt. No. 342. To briefly reiterate, mere references to methods of downloading and distributing copyrighted works (including Plaintiffs’),
and identification of pirated datasets, in no way harm Meta’s competitive position. Instead, Meta seeks to seal them because their disclosure “may lead to [Meta]’s embarrassment,” which is not sufficient to justify sealing. E.g., Kamakana v. City and County of Honolulu, 447 F.3d 1172, 1179 (9th Cir. 2006).Kadrey plaintiffs brief