AI

The Association of American Publishers’ annual meeting on Thursday honored two 2026 anniversaries: the founding of the United States 250 years ago and the passing of the Copyright Act in 1976. At the heart of both, speakers said, is the importance of protecting human expression. At the top of the meeting, AAP ceo Maria Pallante discussed the ongoing litigation related to using pirated books for AI training. Today, the AAP filed a motion for default judgment in their case against pirate site Anna’s Archive, since the site has not responded to the suit. “If the judge grants our motion, we […]

A US district court judge ruled that authors’ suit against Nvidia for copyright infringement to train the chipmaker’s large language models will be allowed to continue. The judge granted part of Nvidia’s motion to dismiss. The lawsuit hinges on three main claims, two of which Judge Jon S. Tigar agreed to let play out in the court: direct, contributory, and vicarious copyright infringement. Authors assert that Nvidia trained the LLMs in its “Megatron family” on the dataset The Pile, which includes Books3, a collection of more than 196,000 pirated books. Nvidia claims that Megatron was trained on “portions of The […]

After multiple lawsuits brought by coalitions of authors against AI companies, now a coalition of book publishers has brought a suit as well. Five publishers, joined by one author, have filed a lawsuit in New York’s Southern District seeking class action status against Meta and ceo Mark Zuckerberg personally for infringing copyrights to train Meta’s LLM Llama. Elsevier, Cengage, Hachette Book Group, Macmillan, McGraw Hill, and Scott Turow claim that Meta, as directed by Zuckerberg, pirated millions of books, scholarly articles, and more, to build and then profit from its AI tool. “Defendants reproduced and distributed millions of copyrighted works […]

Another copyright lawsuit from authors against a tech company for AI training is allowed to continue in Northern California federal court, a judge decided. Plaintiffs Stewart O’Nan, Abdi Nazemian, Brian Keene, Rebecca Makkai, and Jason Reynolds claim that MosaicML, which provides training data for AI companies and is owned by Databricks, used the Books3 dataset to train their own large language model, called MPT, and Databricks’s LLM, called DBRX. Judge Charles B. Dreyer denied the defendants’ motion to dismiss, stating “Defendants may ultimately prevail on this issue, but for now, Plaintiffs’ allegations are sufficient.” The tech company did succeed in getting […]