Anthropic Settles Claude Copyright Case
Anthropic has reached a settlement with the authors whose books it used for training its AI assistant, Claude. The AI developer has agreed to pay the authors $1.5 billion, according to IPLaw360.
The settlement follows a court ruling in June, which I discussed in a previous post. In that opinion, the court found that training AI is fair use, but copyrighted training materials must be purchased, and maintaining a permanent library of the copyrighted works is questionable.
The court’s central finding was that “the purpose and character of using copyrighted works to train LLMs to generate new text [is] quintessentially transformative.” The court also found that the rounds of copying of legitimately purchased books were reasonably necessary for the transformative training use, and the training did not and will not displace the market for copies of the authors’ copyrighted works.
But obtaining pirated copies, as Anthropic did for some of the books, was not a fair use and constituted copyright infringement. The court also held that the separate use of copying the pirated works and building a central library of them was not a fair use.
So this case will not go to trial, with Anthropic electing to pay a sum certain rather than roll the dice on damages. But the court’s June opinion remains a guide for AI developers, i.e., certain training uses may qualify as fair use, but they need to be careful how they acquire and store the copyrighted training materials.