Judge approves AI use of pirated books in a major decision favouring AI firm Anthropic, sparking a shift in how copyright law applies to artificial intelligence. The decision allows Anthropic to train its chatbot, Claude, on millions of books obtained without the consent of the authors. Judge William Alsup decided that this practice is covered by the doctrine of fair use of copyright law in US.
Judge Alsup termed the training of Claude by the company using copyrighted works as being extremely transformative. He said that Anthropic did not sell or directly copy these books, but rather reused them to create a generative AI model in a new and innovative manner.
Court Allows AI Training Using Copyrighted Works
Court reports indicate that Anthropic downloaded pirated books on free sites and digitalised books bought legally. The company aimed to create a complete digital library to facilitate massive AI training. The ruling stated that Anthropic wanted to collect all books in the world to train Claude and other AI systems.
This ruling establishes a possible legal precedent. This case may have an impact on future rulings as musicians, authors, visual artists, and media outlets sue AI companies. The court affirmed that training of AI is a fair use when companies use copyrighted works in a transformative manner.
Stay connected with Closeup Pakistan on X, Medium, Instagram, and Facebook for the latest updates on Entertainment, business & Finance, Shipping, technology, and Current affairs in Pakistan.
