Anthropic, the AI technology company that builds AI services like Claude, has won a court case for using copyrighted content from about 7 million copies of reading material between 2021 and 2024.
The lawsuit was filed last year by Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who filed the suit, alleging that Anthropic had not only pirated their works, but also used the content for machine learning purposes without their knowledge.
Judge William Alsup of the Northern District of California said that the not guilty verdict was given to Anthropic because the copyrighted content used, even without permission, falls under fair use law, where the content used is transformative and there is no outright plagiarism of the content.
While this victory is seen as a major blow to the AI and machine learning industry, Anthropic still faces another charge, namely that it pirated about seven million copies of written content and stored it in its own data centers.
The judge said that even if Anthropic purchased the books that were later used for machine learning, the evidence that they downloaded the pirated content and used it for that purpose was incontrovertible.
He said that another court hearing would be held on the piracy case, and would discuss what appropriate punishment could be imposed on the company.