Anthropic faces another copyright lawsuit, more than 100 authors say it used pirated books

More than 100 authors have sued Anthropic, yet again. The authors accuse the AI company of using copyrighted books without permission to train its Claude.

In Short

• More than 100 authors have filed a copyright lawsuit against Anthropic
• The writers accuse the company of using nearly 500 books without permission
• The lawsuit follows Anthropic’s previous $1.5 billion copyright settlement

AI companies are increasingly operating in a legal grey area as far as copyright is concerned. Result: a number of lawsuits. Now, there is one more in the list as more than 100 authors have filed a lawsuit against Anthropic, accusing it of using their books without permission to train its AI models. The case, filed by Thomas William Shakespeare along with other writers, has once again put spotlight on debate around AI training data.

This is not the first time Anthropic has faced such allegations. In the earlier Bartz v. Anthropic case, authors claimed the company trained its Claude AI model using pirated books. The court later ruled that training AI on legally obtained books could be considered fair use, but found that building a library from pirated copies was a separate issue. The case eventually ended with Anthropic agreeing to pay a $1.5 billion settlement to the affected authors, making it one of the biggest copyright settlements in the US.

Authors accuse Anthropic of using pirated books

The latest lawsuit was filed in the US District Court for the Northern District of California by a group of authors who opted out of the previous settlement.

The authors allege that Anthropic used nearly 500 copyrighted works without authorisation while developing its AI systems. The complaint claims that the company obtained millions of books from sources linked to piracy, including Library Genesis and Pirate Library Mirror, and used them as training material.

The lawsuit names Anthropic CEO Dario Amodei and co-founder Benjamin Mann as defendants. The authors argue that the company benefited from copyrighted material while building its AI technology without compensating the creators behind those works.

Another copyright challenge for Anthropic

The authors are seeking statutory damages of up to $150,000 per infringed work if the court finds the infringement was intentional. Based on the number of works involved, the total claim could reach around $71.4 million.

Notably, the growing tension between AI companies and creators is not limited only to courtrooms. Writers, artists, and other content creators have repeatedly raised concerns that their work is being used as training material without clear consent or compensation. Many argue that while AI systems depend on vast amounts of human-made content to improve, the people behind that content are often left out of the conversation.

This has turned AI training data into one of the biggest debates in the creative industry, with creators demanding more transparency about what goes into building these models.

- Ends