A US judge has ruled that using books to train artificial intelligence (AI) software is not a violation of US copyright law.
In a lawsuit brought against AI firm Anthropic by three writers, a novelist, and two non-fiction authors, the judge has ruled in favor of the AI firm, stating that their use of the authors’ books was “extremely transformative” and therefore allowed under US law.
However, the judge has also stated that the firm will have to stand trial for their use of pirated copies to build their library of material.
Anthropic, which is supported by Amazon and Google’s parent company, Alphabet, could potentially face up to $150,000 in damages per copyrighted work.
The judge noted that the firm currently holds over seven million pirated books in a “central library”.
This ruling is one of the first to address the question of whether Large Language Models (LLMs) can legitimately learn from existing material, a topic that has sparked many legal battles in the industry.
According to Judge Alsup, Anthropic’s LLMs were trained on works in order to create something new, rather than to replicate or replace existing works.
He also stated that the authors did not claim that the training resulted in “infringing knockoffs” of their works.
Similar legal battles have arisen over the use of other media and content by the AI industry, such as journalism articles, music, and video.
Recently, Disney and Universal filed a lawsuit against AI image generator Midjourney for piracy, and the BBC is considering legal action for the unauthorized use of its content.
In response to these legal battles, some AI companies have started striking deals with creators and publishers to obtain proper licenses for their material.
Judge Alsup’s ruling allows for a “fair use” defense in future legal judgements, but also states that Anthropic violated the authors’ rights by saving pirated copies of their books in a “central library of all the books in the world”.
In a statement, Anthropic expressed their satisfaction with the judge’s recognition of their use of the works as transformative, but disagreed with the decision to hold a trial regarding how some of the books were obtained and used.
The company remains confident in their case and is considering their options, while a lawyer for the authors declined to comment.
The authors involved in the case are Andrea Bartz, a best-selling mystery thriller writer, and non-fiction writers Charles Graeber and Kirk Wallace Johnson.