Artificial intelligence startup Anthropic has settled a class action copyright infringement lawsuit, in which authors accused the company of training its AI models on their work without permission, according to a Tuesday court filing.
San Francisco-based Anthropic, which trained its AI assistant Claude using copyrighted books, was sued by authors Andrea Bartz, Charles Graeber and Kirk Wallace in August 2024.
Terms of the settlement were not disclosed. The company did not immediately respond to a request for comment.
Last June, U.S. District Judge William Alsup ruled that AI models could legally learn from copyrighted books without the authors’ consent. The decision was a partial win for Anthropic.
Alsup found the usage to be “exceedingly transformative” and “a fair use,” though the company might have broken the law by pirating a large portion of its source material. According to the filing in the U.S. Court of Appeals for the 9th Circuit, the tech company and the involved authors asked the court to pause its proceedings while they finalize the settlement deal.
“Fair use” doctrine, which allows for the limited reproduction of copyrighted material without consent in some circumstances, is a key component of AI companies’ defenses against copyright claims.
“This historic settlement will benefit all class members,” said Justin Nelson, partner at Susman Godfrey and lawyer for the authors in a statement to the Times. “We look forward to announcing details of the settlement in the coming weeks.”
Alsup originally ordered the matter to go to trial in December to decide how much they would pay in piracy damages. If it went to trial, the damages could have reached up to $150,000 per case of willful copyright infringement and could have cost the startup billions. In early August, the AI company attempted to get an appeal and was denied.
Anthropic, founded by former OpenAI employees and backed by Amazon, pirated at least 7 million books from Books3, Library Genesis and Pirate Library Mirror, online libraries containing unauthorized copies of copyrighted books, to train its software, according to the judge.
They also bought millions of print copies in bulk and stripped the books’ bindings, cut their pages and scanned them into digital and machine-readable forms, which Alsup found to be in the bounds of fair use, according to the judge’s ruling.
“The mere conversion of a print book to a digital file to save space and enable searchability was transformative for that reason alone,” he wrote.
Anthropic later purchased the books it initially pirated. Alsup said the purchases did not absolve the company, but that they could reduce damages.
This agreement is being reached as many other copyright cases against AI companies are being brought to courts around the country. Most recently, Walt Disney Co. and Universal Pictures sued artificial intelligence company Midjourney, which the studios allege trained its image generation models on their copyrighted materials.
Artificial intelligence startup Anthropic has settled a class action copyright infringement lawsuit, in which authors accused the company of training its AI models on their work without permission, according to a Tuesday court filing.
San Francisco-based Anthropic, which trained its AI assistant Claude using copyrighted books, was sued by authors Andrea Bartz, Charles Graeber and Kirk Wallace in August 2024.
Terms of the settlement were not disclosed. The company did not immediately respond to a request for comment.
Last June, U.S. District Judge William Alsup ruled that AI models could legally learn from copyrighted books without the authors’ consent. The decision was a partial win for Anthropic.
Alsup found the usage to be “exceedingly transformative” and “a fair use,” though the company might have broken the law by pirating a large portion of its source material. According to the filing in the U.S. Court of Appeals for the 9th Circuit, the tech company and the involved authors asked the court to pause its proceedings while they finalize the settlement deal.
“Fair use” doctrine, which allows for the limited reproduction of copyrighted material without consent in some circumstances, is a key component of AI companies’ defenses against copyright claims.
“This historic settlement will benefit all class members,” said Justin Nelson, partner at Susman Godfrey and lawyer for the authors in a statement to the Times. “We look forward to announcing details of the settlement in the coming weeks.”
Alsup originally ordered the matter to go to trial in December to decide how much they would pay in piracy damages. If it went to trial, the damages could have reached up to $150,000 per case of willful copyright infringement and could have cost the startup billions. In early August, the AI company attempted to get an appeal and was denied.
Anthropic, founded by former OpenAI employees and backed by Amazon, pirated at least 7 million books from Books3, Library Genesis and Pirate Library Mirror, online libraries containing unauthorized copies of copyrighted books, to train its software, according to the judge.
They also bought millions of print copies in bulk and stripped the books’ bindings, cut their pages and scanned them into digital and machine-readable forms, which Alsup found to be in the bounds of fair use, according to the judge’s ruling.
“The mere conversion of a print book to a digital file to save space and enable searchability was transformative for that reason alone,” he wrote.
Anthropic later purchased the books it initially pirated. Alsup said the purchases did not absolve the company, but that they could reduce damages.
This agreement is being reached as many other copyright cases against AI companies are being brought to courts around the country. Most recently, Walt Disney Co. and Universal Pictures sued artificial intelligence company Midjourney, which the studios allege trained its image generation models on their copyrighted materials.
Artificial intelligence startup Anthropic has settled a class action copyright infringement lawsuit, in which authors accused the company of training its AI models on their work without permission, according to a Tuesday court filing.
San Francisco-based Anthropic, which trained its AI assistant Claude using copyrighted books, was sued by authors Andrea Bartz, Charles Graeber and Kirk Wallace in August 2024.
Terms of the settlement were not disclosed. The company did not immediately respond to a request for comment.
Last June, U.S. District Judge William Alsup ruled that AI models could legally learn from copyrighted books without the authors’ consent. The decision was a partial win for Anthropic.
Alsup found the usage to be “exceedingly transformative” and “a fair use,” though the company might have broken the law by pirating a large portion of its source material. According to the filing in the U.S. Court of Appeals for the 9th Circuit, the tech company and the involved authors asked the court to pause its proceedings while they finalize the settlement deal.
“Fair use” doctrine, which allows for the limited reproduction of copyrighted material without consent in some circumstances, is a key component of AI companies’ defenses against copyright claims.
“This historic settlement will benefit all class members,” said Justin Nelson, partner at Susman Godfrey and lawyer for the authors in a statement to the Times. “We look forward to announcing details of the settlement in the coming weeks.”
Alsup originally ordered the matter to go to trial in December to decide how much they would pay in piracy damages. If it went to trial, the damages could have reached up to $150,000 per case of willful copyright infringement and could have cost the startup billions. In early August, the AI company attempted to get an appeal and was denied.
Anthropic, founded by former OpenAI employees and backed by Amazon, pirated at least 7 million books from Books3, Library Genesis and Pirate Library Mirror, online libraries containing unauthorized copies of copyrighted books, to train its software, according to the judge.
They also bought millions of print copies in bulk and stripped the books’ bindings, cut their pages and scanned them into digital and machine-readable forms, which Alsup found to be in the bounds of fair use, according to the judge’s ruling.
“The mere conversion of a print book to a digital file to save space and enable searchability was transformative for that reason alone,” he wrote.
Anthropic later purchased the books it initially pirated. Alsup said the purchases did not absolve the company, but that they could reduce damages.
This agreement is being reached as many other copyright cases against AI companies are being brought to courts around the country. Most recently, Walt Disney Co. and Universal Pictures sued artificial intelligence company Midjourney, which the studios allege trained its image generation models on their copyrighted materials.
Artificial intelligence startup Anthropic has settled a class action copyright infringement lawsuit, in which authors accused the company of training its AI models on their work without permission, according to a Tuesday court filing.
San Francisco-based Anthropic, which trained its AI assistant Claude using copyrighted books, was sued by authors Andrea Bartz, Charles Graeber and Kirk Wallace in August 2024.
Terms of the settlement were not disclosed. The company did not immediately respond to a request for comment.
Last June, U.S. District Judge William Alsup ruled that AI models could legally learn from copyrighted books without the authors’ consent. The decision was a partial win for Anthropic.
Alsup found the usage to be “exceedingly transformative” and “a fair use,” though the company might have broken the law by pirating a large portion of its source material. According to the filing in the U.S. Court of Appeals for the 9th Circuit, the tech company and the involved authors asked the court to pause its proceedings while they finalize the settlement deal.
“Fair use” doctrine, which allows for the limited reproduction of copyrighted material without consent in some circumstances, is a key component of AI companies’ defenses against copyright claims.
“This historic settlement will benefit all class members,” said Justin Nelson, partner at Susman Godfrey and lawyer for the authors in a statement to the Times. “We look forward to announcing details of the settlement in the coming weeks.”
Alsup originally ordered the matter to go to trial in December to decide how much they would pay in piracy damages. If it went to trial, the damages could have reached up to $150,000 per case of willful copyright infringement and could have cost the startup billions. In early August, the AI company attempted to get an appeal and was denied.
Anthropic, founded by former OpenAI employees and backed by Amazon, pirated at least 7 million books from Books3, Library Genesis and Pirate Library Mirror, online libraries containing unauthorized copies of copyrighted books, to train its software, according to the judge.
They also bought millions of print copies in bulk and stripped the books’ bindings, cut their pages and scanned them into digital and machine-readable forms, which Alsup found to be in the bounds of fair use, according to the judge’s ruling.
“The mere conversion of a print book to a digital file to save space and enable searchability was transformative for that reason alone,” he wrote.
Anthropic later purchased the books it initially pirated. Alsup said the purchases did not absolve the company, but that they could reduce damages.
This agreement is being reached as many other copyright cases against AI companies are being brought to courts around the country. Most recently, Walt Disney Co. and Universal Pictures sued artificial intelligence company Midjourney, which the studios allege trained its image generation models on their copyrighted materials.
