Anthropic won a stern victory in the ongoing legal battle about artificial intelligence models and copyright, which may resound in dozens of other AI copyrights that flow through the legal system in the United States. The court determined that the training of AI tools in the field protected by copyright is anthropic, arguing that the behavior is protected by the doctrine of “allowed use”, which allows unauthorized exploit of copyrights protected by copyright under certain conditions.
“The use of the training was allowed”, senior district judge William Alsup wrote In the summary order of the sentence issued delayed Monday evening. In the copyright, one of the main ways in which the courts determine whether the exploit of copyrights protected by copyright without permission is allowed to be used, it is to examine whether the exploit was “transformative”, which means that it is not a substitute for the original work, but rather something novel. “The disputed technology was one of the most transformative of us that it will see in our lives,” wrote ALSUP.
“This is the first serious ruling in a generative case regarding AI copyrights to deal with allowed use in detail,” says Chris Mammen, managing partner at Wombble Bond Dickinson, who focuses on intellectual property law. “Judge ALSUP stated that LLM training has a transformation application – even when there is a significant remembering. In particular, he rejected the argument that what people do while reading and remembering, differs from what computers do during LLM training.”
The case, a collective lawsuit filed by the authors of the book, who claimed that Anthropic violated his copyright by means of his work without permission, was first submitted in August 2024 in the American District Court for the Northern California district.
Anthropic is the first artificial intelligence company that won this type of battle, but the victory has a substantial star. While ALSUP stated that Anthropik’s training was allowed, it ruled that the authors could take anthropic attempt over the piracy of their works.
While Antropic finally went to training in the purchased copies of books, yet he first collected and maintained a huge library of pirate materials. “Anthropic downloaded over seven million pirate copies of books, paid nothing and held these pirate copies in his library, even after making the decision that she would not use them to train her artificial intelligence (in general or never again). The authors argue that Anthropian should pay for these pirate library copies. The prohibition is agreed,” writes Alsup.
“We will have a test on pirate copies used to create the Central Library of Anthropik and the resulting damage,” he sums up the order.
Anthropic did not immediately answer the requests for comment. Lawyers of the plaintiffs refused to comment.
Lawsuit, Bartz against Anthropic He was first submitted less than a year ago; Anthropic asked for a summary sentence regarding the permitted exploit in February. It is worth noting that ALSUP has much more experience with questions allowed than the average federal judge, because it presided over the initial process in Google against OracleA breakthrough case regarding technology and copyright, which eventually passed before the Supreme Court.