Major media, music, and technology players are deepening their legal battles over generative artificial intelligence systems, with courts now focused on how training data was obtained, how outputs were generated, and how traditional copyright and contract doctrines apply. Sony Music Entertainment is pressing to preserve its key claim under the Digital Millennium Copyright Act against Uncharted Labs, arguing that YouTube’s “rolling cipher” qualifies as an unlawful access-control circumvention method in light of a recent California ruling in Cordova v. Huneault. Udio counters that the Cordova decision is unpersuasive and insists the technology functions only as a copy-control system, which it says falls outside the DMCA’s anticircumvention scope.
In the Southern District of New York’s multi-district litigation involving the New York Times and Daily News, fact discovery closed on February 26 but disputes continue over the attempted deposition of professor Tom Goldstein, a computer science expert who helped generate some Artificial Intelligence outputs cited as evidence of infringement. Plaintiffs seek a protective order, asserting work-product protection that the court has previously upheld, while defendants argue that privilege was waived by court-ordered technology presentations describing output-generation methods. A joint stipulation filed on the discovery deadline confirmed that at least nine February depositions of former OpenAI executives Dario Amodei, Mira Murati, Ilya Sutskever, current OpenAI employees, and Boston Consulting Group were moved into March, as the case transitions into expert discovery.
Entertainment studios are also pushing ahead against non-us Artificial Intelligence companies. Disney, Universal, and Warner Bros. have now completed service on MiniMax, Shanghai Xiyu Jizhi Technology Co. Ltd., and Nanonoble Pte. Ltd., with Quinn Emanuel accepting service effective February 20, 2026 and defendants’ response to the complaint due by April 24, 2026, while reserving challenges to jurisdiction, venue, and sufficiency. In Sarah Andersen v. Stability AI, the court granted leave to amend, allowing new direct infringement allegations tied to training datasets for DeviantArt’s DreamUp and Midjourney’s use of the Datacomp xLarge dataset, which marks a notable procedural win for artists. In Richard Kadrey v. Meta and UMG Recordings v. Suno, there were no new substantive rulings, but in the Suno matter the parties jointly asked to move the fact discovery deadline from March 6 to June 26 to accommodate ongoing discovery disputes and related schedule adjustments for expert work and dispositive motions.
Music publishers and platforms face parallel discovery and jurisdiction fights in cases against Anthropic and Reddit. In Concord Music Group v. Anthropic, the defendant is seeking relief from a magistrate’s split order on investigative prompts and outputs, arguing it needs both pre-suit and post-suit prompts to separate plaintiffs’ crafted queries from third-party use, and the district court has now extended deadlines and moved trial to January 25, 2027. In Reddit v. Anthropic, the court has asked both sides to brief whether Reddit’s content falls under 17 U.S.C. §§ 102-103 and whether non-copyrightable data such as usernames, vote counts, and comment nesting can sustain common law and contract claims, while Anthropic disputes both the scope of copyright protection and the existence of an enforceable browsewrap agreement. Related Disney, Universal, and Warner Bros. litigation against Midjourney remains mostly in early scheduling, with a mediation report due August 21, 2026, non-expert discovery closing September 21, 2026, and expert discovery closing November 9, 2026. Meanwhile, the Hendrix v. Apple class action has been consolidated with other author suits and now has interim co-lead counsel appointed through January 2027, with an amended complaint detailing how Apple’s integrated models and generative Artificial Intelligence outputs allegedly impair markets for human-created works, setting up future fair use and market-harm arguments.
