Anthropic wins key ruling on AI in authors copyright lawsuit
(Reuters) -San Federal Judge in Francisco decided that the use of an unauthorized book to train anthropic’s artificial intelligence system late on Monday was legal.
An important question for the AI industry, the US regional judge William Alsup, who is engaged in technology companies, said anthropic’s writers Andrea Bartz, Charles Graeber and Kirk Wallace Johnson to educate the big language model.
However, Alsup also said that the storage of Anthropic’s books “Central Library” violated the copyrights and is not a fair use.
For the anthropic spokesman and writers, lawyers did not immediately respond to their requests for comments on Tuesday.
The authors filed anthropic lawsuits last year, arguing that the company, which was supported by Amazon and Alphabe, used pirate versions of their books without permission or compensation to teach Claude to respond to human requests.
Class Action Case is one of the few people brought by authors, news organizations and other copyright holders against companies, including Openai, Microsoft and commodity platforms on AI trainings.
The fair usage doctrine allows the use of copyrights to be used in some cases without permission of the copyright owner.
Fair use is an important legal defense for technology companies, and Alsup’s decision is the first decision in the context of productive artificial intelligence.
AI companies argue that their systems use the materials that are protected with copyrights to create new, transformative content in a fair way and that they are forced to pay to copyright owners for their jobs, and that they can hamstring the developing AI industry.
He encouraged human creativity, because he used books in a fair way to the anthropic court, and “not only permits, but also encourages the US copyright law. The company said that the books of the system copied “plaintiffs’ spelling, to use uninterrupted information and to use what they have learned to create revolutionary technology”.
Copyright holders say that AI companies have illegally copying their work to produce competitors that threaten their livelihoods.
On Monday, Alsup admitted that his education was “extremely transformative”.
Alsup, “Like any reader who wants to be a writer, like anthropic’s LLMs, in progress and to copy or support them – but trained to work to turn a hard corner and create something different.” He said.
(Reporting by Blake Britain in Washington; Organization by Chizu Nomiyama and Louise Heavens)
