A US judge determined that tech companies were using books to train artificial intelligence systems without the author's permission.
A federal judge in San Francisco said that humanity will “use fair use” the book by authors Andrea Burtz, Charles Graeber and Kirk Wallace Johnson to train its Claude-Borz.
Judge William Alsup compared the use of humanity's books with “readers aiming to become writers,” who use works that “turn hard corners and create something different,” rather than “don't race or recreate them first.”
However, Alsup added that humanity's copying and storage of pirated books over 7m in the Central Library had infringed the author's copyright and was not using them fairly. The judge ordered a trial in December to determine how much humanity is in the infringement.
“The fact that humanity later purchased a copy of a book that he had stolen from the Internet will not be liable for theft, but it could affect the extent of statutory damage,” writes Alsup.
According to US copyright law, intentional copyright infringement could result in $150,000 (£110,000) of damage.
Copyright issues have pitted AI companies against publishers and creative industries. This is because the Generated AI Model (the term for technology that underpins powerful tools such as ChatGPT chatbots) requires training on a huge amount of public data to generate responses. Much of that data includes copyrighted works.
A human spokesman said the company was pleased that the court recognized that AI training was transformative and “consistent with copyright objectives that enable creativity and promote scientific advancement.”
John Strand, a copyright lawyer at the US law firm Wolf Greenfield, said decisions from “respected” judges were “very important.”
He added: “Dozens of other cases are pending across the United States, including similar questions about copyright infringement and fair use. The decision of Justice Alsup here is something that other courts must consider in their own cases.”
Strand said that as other AI copyright cases have passed the legal system:
The writers filed a proposed class action lawsuit against humanity last year, claiming that Amazon and Alphabet-backed companies would use pirated versions of their books without permission or compensation to teach Claude to respond to human prompts.
The proposed class action lawsuit is one of several lawsuits filed by authors, news outlets and other copyright holders against companies such as Openai, Microsoft and the Meta platform over AI training.
Fair use doctrine allows copyrighted works to be used in some circumstances without permission from the copyright owner. Fair use is an important legal defense for high-tech companies, and the Allsup decision is the first decision to address it in the context of generating AI.
After the newsletter promotion
AI companies argue that systems could smash early industries when they are forced to use copyrighted material fairly to create new, transformative content and pay copyright owners for their work. Humanity told the court that the book was fair and that US copyright law “not only allows but encourages AI training to promote human creativity.”
The company said the system copied the book to “study the plaintiff's writings, extract undeveloped information from them, and use what they learned to create innovative technologies.”
Giles Parsons, a partner at British law firm Brown Jacobson, said the ruling would not have any impact in the UK, where fair use debates are more shaky. Under current UK copyright laws, which the government is about to change, copyrighted work can be used without permission from scientific or academic research.
He said: “The UK has a much narrower fair use defense that is very unlikely to apply in these circumstances.”
Copyright holders in the US and UK say AI companies are illegally copying their work and generating competing content that threatens their livelihoods. The UK government's proposal to change UK copyright laws by using copyrighted work without permission is met with noisy opposition from the creative industry, unless the work owner wants to opt out of the process.
Alsup said that humanity violated the author's rights by storing pirated copies of their books as part of the “central library of all books in the world.” Humanity and other prominent AI companies, including Openai and Facebook owner Meta, have been accused of downloading pirated copies of millions of books to train their systems.
