TLDR: AI startup Anthropic has settled a class-action lawsuit filed by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who accused the company of using their copyrighted books to train its AI assistant, Claude, without permission. While a judge previously ruled that training AI models on copyrighted material could be considered ‘fair use,’ Anthropic was still facing a trial over its alleged piracy and storage of up to 7 million books. The terms of the settlement, announced on August 26, 2025, remain undisclosed, but legal representatives for the authors stated it would benefit all class members.
San Francisco-based artificial intelligence firm Anthropic, known for its AI assistant Claude and backed by tech giants Amazon and Alphabet, has reached a settlement in a significant class-action copyright infringement lawsuit. The agreement, disclosed in a federal appeals court filing on Tuesday, August 26, 2025, marks the first resolution in a series of high-profile legal battles concerning AI training and intellectual property in the United States.
The lawsuit, originally filed in August 2024 by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, alleged that Anthropic engaged in ‘large-scale theft’ by utilizing their copyrighted literary works to train its AI models without authorization or compensation.
This development follows a complex ruling in June by U.S. District Judge William Alsup of San Francisco. In a partial victory for Anthropic, Judge Alsup determined that AI models could legally learn from copyrighted books without explicit author consent, deeming such use ‘exceedingly transformative’ and falling under the ‘fair use’ doctrine.
However, the judge also found that Anthropic might have violated copyright law by pirating and storing a vast collection of up to 7 million books. These materials were reportedly sourced from ‘shadow libraries’ such as Books3, Library Genesis, and Pirate Library Mirror. Judge Alsup clarified that while the training aspect might qualify as fair use, the copying and storage of these pirated books in a ‘central library’ did not and constituted copyright infringement.
Consequently, Anthropic was scheduled to face a trial in December to determine the extent of damages owed for this alleged piracy. Under U.S. copyright law, willful infringement can lead to statutory damages of up to $150,000 per work.
Also Read:
- Japanese Media Giants File Landmark Copyright Lawsuit Against AI Firm Perplexity
- Perplexity Navigates AI Data Landscape with Publisher Revenue Share Amidst Scraping Allegations and Copyright Suits
The specific terms of the newly announced settlement have not been disclosed, and an Anthropic spokesperson declined to comment on the filing. However, Justin Nelson, an attorney representing the authors, issued a statement affirming that ‘This historic settlement will benefit all class members.’ This resolution is expected to influence future legal disputes between AI companies and content creators, potentially setting a precedent for how such cases are handled.
