Federal Judge William Alsup of the Northern District of California has ruled that Anthropic's practice of training artificial intelligence models on purchased copies of books without the authors' permission is legal.
According to TechCrunch, this decision was a significant victory for Anthropic in its copyright infringement case, as well as a blow to authors, artists, and publishers who had filed dozens of lawsuits against OpenAI, Meta, Midjourney, Google, and others.
It is noted that these lawsuits often depend on how a judge interprets the fair use doctrine — a passage from copyright law that has not been updated since 1976 — from a time before the Internet, let alone the concept of training generative artificial intelligence.
When making decisions about fair use, the purposes for which the work is used are taken into account—parody and education may be acceptable here.
In this particular case, Bartz v. Anthropic, a group of plaintiff authors challenged the way Anthropic obtained and stored their works. According to the lawsuit, Anthropic sought to create a "central library" of "all the books in the world" for "perpetual preservation." But millions of these copyrighted books were being downloaded for free from pirate sites, which is clearly illegal.
While the judge found that Anthropic's training on these materials was in good faith, the court will consider the nature of this so-called "central library."
"We will hold a trial regarding the pirated copies used to create Anthropic's central library and the damages caused. The fact that Anthropic later purchased a copy of a book that it had previously stolen from the Internet does not absolve it of liability for the theft, but it may affect the amount of damages," the judge wrote in the decision.
As a reminder, Anthropic recently launched a new section on its website called Claude Explains — a blog mostly written by the Claude language model. The page publishes materials on the use of AI in various areas — from code analysis to writing business strategies.
Read also: Anthropic CEO says AI will take half of office jobs and increase unemployment by 10-20%