AI copyright cases heat up
A new wave of copyright and fair‑use litigation is pressuring AI training pipelines — courts are scrutinizing large‑scale scraping and a recent ruling about a company that ignored 163,000 notices shows the legal terrain is changing fast. (traverselegal.com, smithstephen.com)
The U.S. Supreme Court issued a unanimous 9–0 opinion in Cox Communications v. Sony Music on March 25, 2026, holding that contributory copyright liability requires intent to induce infringement or a service designed for infringement. (supreme.justia.com) The Court vacated the roughly $1 billion jury judgment the Fourth Circuit had affirmed and said evidence that an ISP received infringement notices was insufficient, absent intent, to establish secondary liability. (scotusblog.com) Federal courts remain fractured on AI training fair use: in Thomson Reuters v. ROSS Intelligence (D. Del., Feb. 11, 2025) the court rejected the defendant’s fair‑use defense for training a legal‑research AI and found market‑harm to Westlaw. (dwt.com) By contrast, in Bartz v. Anthropic (N.D. Cal., June 23, 2025) Judge William Alsup granted summary judgment that training an LLM on lawfully purchased books could be “transformative” and therefore fair use while leaving alleged use of pirated copies for trial. (cnbc.com) BakerHostetler’s AI/copyright case tracker lists multiple active suits over model training and dataset sourcing—naming cases against Cohere, Stability AI, Anthropic, GitHub/Microsoft/OpenAI, and a Stability AI trial currently calendared for April 5, 2027. (bakerlaw.com) Platform and publisher enforcement is already escalating: Reddit filed an “industrial‑scale” scraping suit in October 2025 against Perplexity and three data‑extraction firms, and publishers plus Big Tech are pursuing direct licensing and technical blocking measures. (pbs.org) Traverse Legal’s roundup highlights that courts are actively testing how “transformation” and the fourth‑factor market‑harm inquiry apply to model training, producing divergent rulings that leave large‑scale scraping policies unclear for developers and platforms. (traverselegal.com) Commentators warn the Cox decision narrows remedies tied to notice‑and‑takedown regimes, pushing rights holders to rely more on discovery into model datasets and on evidence of competitive market harm when suing AI trainers. (musictechpolicy.com)
