In a pivotal ruling shaping the future of artificial intelligence and copyright law, a US judge has declared that training AI models with copyrighted books constitutes “transformative use” and does not violate US copyright law, according to a bbc.co.uk report. However, the AI firm Anthropic, backed by tech giants Amazon and Alphabet, will still face trial over its use of pirated copies to build its vast digital library.
The decision came in a lawsuit filed last year by authors Andrea Bartz, Charles Graeber and Kirk Wallace Johnson, who accused Anthropic of using their works to train its Claude AI model. Judge William Alsup stated that Anthropic’s use of the books was “exceedingly transformative,” emphasising that the AI learned from the works to “create something different,” not to replicate them.
Despite this win for the AI industry, Judge Alsup rejected Anthropic’s motion to dismiss the case. He highlighted that Anthropic holds over seven million pirated books in a “central library,” a practice that violates the authors’ rights. The firm could face damages of up to $150,000 per copyrighted work.
This ruling is among the first to address how Large Language Models (LLMs) can legitimately learn from existing material, a question at the heart of numerous legal battles across various media. While acknowledging the transformative nature of AI training, the judge’s decision underscores the critical distinction between fair use and the illicit acquisition of copyrighted material.
•Featured Image: Andrea Bartz, one of a number of writers who have taken legal action over AI/Getty Images
