Humanity, an artificial intelligence company, has recorded a major victory in court that could affect dozens of similar cases. Despite a legal victory, the company still faces claims of copyright infringement that allegedly pirated the book to launch its digital library.
Fair Use Discussion
The author group filed a class action lawsuit against humanity in 2024, claiming that they used the book without permission to violate copyright.
Download the SAN app now to stay fair and up to date. Straight Facts™.
Point phone camera here
The company used these works to train a language model called Claude, a competitor to popular AI tools such as Openai's ChatGpt and Google's Gemini.
Humanity argued for fair use that could limit the use of copyrighted material without allowing news reporting, education or other purposes.
One of the main ways that courts can determine fair use is to see if the use of copyrighted works is “transformative.” In other words, it's not a replacement for the original work, it's something new.
“The technology in question was one of the most transformative things that many of us could see in our lifetimes,” Judge William Alsup wrote in his summary judgment. “The use of problematic books to train Claude and its precursors was very transformative and was a fair use under Section 107 of the Copyright Act.”
Decision to split
Despite the ruling in favour of the fair use of Allsup's humanity, the judge also ruled that copyright infringement of the book is not legally permissible.
What that basically means is that although it is not permitted to download and build libraries from pirated books of humanity, it is to train AI tools using books.
“Before buying books for the Central Library, humanity downloaded more than 7 million copies of pirated copies, kept these pirated copies in the library even after it decided it would not pay anything and not use them to train AI,” wrote Judge Alsup. “The downloaded pirated copies used to construct the Central Library were not justified by fair use.”
Trials are now being held regarding allegations of copyright infringement. The minimum statutory loss for this type of copyright infringement is $750 per book.
The judge's summary states that humanity has pirated seven million books. This means that losses can exceed up to $5 billion for the company.
A recent report from Reuters shows that humanity is making annual revenues of around $3 billion.
AI lawsuit
This is far from the first lawsuit involving AI and fair use, but it is one of the first successful fair use debates.
In another lawsuit, Thomson Reuters sued AI startup Loss Intelligence, claiming that Ross used Westlow's headnotes to train Ross' AI legal research engine. Westlaw is a renowned legal research platform, a subsidiary of Reuters.
Ross tried to assert fair use, but the court said some of the fair use factors had done against Ross. The court said that Ross was using Westrow's prime minister, which undermined Westrow and its derivative products markets.
Now, one ruling opposes fair use and this latest ruling is in favour of fair use, so it could be appropriate for the U.S. Supreme Court to consider it.
In a recent filing, Disney and Universal filed a lawsuit against the Mid Journey. The Hollywood powerhouse claims that Midjourney trained AI image generators using copyrighted works that include images such as Marvel, The Simpsons, and more.
Midjourney has not commented on the lawsuit yet. Another lawsuit was filed by Getty Images, filed against Stability AI, who is allegedly using Getty images to train Stability AI tools that use Getty images to generate images from text input.
Cole Lauterbach (Managing Editor)
and Devin Pavlou (digital producer)
I contributed to this report.
