Anthropic Secures Landmark Fair Use Victory Amid Claims of Book Copyright Infringement

A federal judge has ruled in favor of Anthropic in a key AI copyright case, determining that using legally acquired books to train its AI models constitutes fair use, even without explicit author permission. This landmark decision, a first for the AI industry, is narrowly focused on the specific act of purchasing physical books and converting them into digital training data.
However, Judge William Alsup of the Northern District of California clarified in his ruling that Anthropic must face a separate trial for allegedly pirating "millions" of books online. The decision also leaves open the question of whether AI-generated outputs infringe copyright, a central issue in other pending lawsuits.
The case was brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who sued Anthropic last year, accusing the company of training its Claude AI models on unlawfully obtained material. This pivotal ruling may influence how future AI copyright cases are adjudicated.
The judge specifically addressed Anthropic's process of buying printed books, removing their bindings, cutting the pages, and scanning them into a central digital library for AI training. He ruled that both the digitization of legally purchased books and their subsequent use for training a large language model (LLM) were transformative enough to qualify as fair use.
“The authors' complaint is akin to arguing that teaching schoolchildren to write well would lead to a flood of competing works,” Judge Alsup wrote. He added that the Copyright Act “aims to promote the creation of original works, not to shield authors from competition.”
Despite these favorable findings for Anthropic, Judge Alsup stated that the company's storage of millions of pirated book copies in its library—regardless of whether all were used for training—does not constitute fair use. “This order doubts any defendant could justify why downloading source copies from pirate sites when lawful alternatives were available was reasonably necessary for a subsequent fair use,” Alsup emphasized.
The court will conduct a separate trial concerning the pirated content used by Anthropic to determine any resulting damages.
“We are gratified the Court acknowledged that using 'works to train LLMs was transformative — spectacularly so,'” Anthropic spokesperson Jennifer Martinez said in an emailed statement to The Verge. “Aligning with copyright's goal of fostering creativity and scientific progress, 'Anthropic's LLMs were trained not to replicate or replace existing works, but to take a new direction and create something entirely different.'”
Update, June 24th: Added a statement from Anthropic.
Related article
Google Photos brings Clueless's iconic closet to life with AI
Google Photos announced a new AI-powered feature on Wednesday that will soon turn photos of your clothes into a digital closet, letting you create fresh outfit combinations and even virtually try them on. The concept clearly draws inspiration from Ch
Notion transforms its workspace into a hub for AI agents
Notion, the productivity software company, is entering the agentic era.During a live-streamed product announcement on Wednesday, Notion—best known for its collaborative note-taking app—unveiled a new developer platform that extends the capabilities o
ElevenLabs names BlackRock, Jamie Foxx, Eva Longoria as new investors
ElevenLabs, the voice AI company, has disclosed additional investors in its $500 million Series D round, originally announced in February. These include institutional investors like BlackRock, Wellington, D.E. Shaw, and Schroders; corporations such a
Related Special Topic Recommendations
Comments (0)
0/500

A federal judge has ruled in favor of Anthropic in a key AI copyright case, determining that using legally acquired books to train its AI models constitutes fair use, even without explicit author permission. This landmark decision, a first for the AI industry, is narrowly focused on the specific act of purchasing physical books and converting them into digital training data.
However, Judge William Alsup of the Northern District of California clarified in his ruling that Anthropic must face a separate trial for allegedly pirating "millions" of books online. The decision also leaves open the question of whether AI-generated outputs infringe copyright, a central issue in other pending lawsuits.
The case was brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who sued Anthropic last year, accusing the company of training its Claude AI models on unlawfully obtained material. This pivotal ruling may influence how future AI copyright cases are adjudicated.
The judge specifically addressed Anthropic's process of buying printed books, removing their bindings, cutting the pages, and scanning them into a central digital library for AI training. He ruled that both the digitization of legally purchased books and their subsequent use for training a large language model (LLM) were transformative enough to qualify as fair use.
“The authors' complaint is akin to arguing that teaching schoolchildren to write well would lead to a flood of competing works,” Judge Alsup wrote. He added that the Copyright Act “aims to promote the creation of original works, not to shield authors from competition.”
Despite these favorable findings for Anthropic, Judge Alsup stated that the company's storage of millions of pirated book copies in its library—regardless of whether all were used for training—does not constitute fair use. “This order doubts any defendant could justify why downloading source copies from pirate sites when lawful alternatives were available was reasonably necessary for a subsequent fair use,” Alsup emphasized.
The court will conduct a separate trial concerning the pirated content used by Anthropic to determine any resulting damages.
“We are gratified the Court acknowledged that using 'works to train LLMs was transformative — spectacularly so,'” Anthropic spokesperson Jennifer Martinez said in an emailed statement to The Verge. “Aligning with copyright's goal of fostering creativity and scientific progress, 'Anthropic's LLMs were trained not to replicate or replace existing works, but to take a new direction and create something entirely different.'”
Update, June 24th: Added a statement from Anthropic.
Google Photos brings Clueless's iconic closet to life with AI
Google Photos announced a new AI-powered feature on Wednesday that will soon turn photos of your clothes into a digital closet, letting you create fresh outfit combinations and even virtually try them on. The concept clearly draws inspiration from Ch
Notion transforms its workspace into a hub for AI agents
Notion, the productivity software company, is entering the agentic era.During a live-streamed product announcement on Wednesday, Notion—best known for its collaborative note-taking app—unveiled a new developer platform that extends the capabilities o
ElevenLabs names BlackRock, Jamie Foxx, Eva Longoria as new investors
ElevenLabs, the voice AI company, has disclosed additional investors in its $500 million Series D round, originally announced in February. These include institutional investors like BlackRock, Wellington, D.E. Shaw, and Schroders; corporations such a





Home






