Anthropic to pay over 220 billion yen to settle copyright infringement lawsuit
Anthropic, the developer of the chat AI 'Claude,' has been sued by a group of authors for copyright infringement, and it has been revealed that the company will pay $1.5 billion (approximately 220 billion yen) to settle the case.
“First of its kind” AI settlement: Anthropic to pay authors $1.5 billion - Ars Technica
In August 2024, authors Andrea Bartz, Charles Graeber, and Kirk Johnson sued Anthropic, alleging copyright infringement by using data from pirate sites like LibGen and Books3, as well as scans of physical books, to train Claude.
Anthropic sued by three authors for copyright infringement, claiming it used hundreds of thousands of copyrighted books in training for Claude - GIGAZINE
The U.S. District Court for the Northern District of California, which heard the case, subsequently ruled that training an AI with a legally purchased book, even without the author's permission, constitutes fair use and does not constitute copyright infringement.
A ruling states that AI companies don't need permission from the author to train their AI with legally acquired books.
However, documents filed with the court in late August 2025 revealed that Anthropic and the artists' group had agreed to a settlement. It was pointed out that Anthropic agreed to the settlement in order to avoid paying high damages.
Anthropic settles with author group over AI learning copyright infringement lawsuit, avoiding damages of up to 150 trillion yen - GIGAZINE
Furthermore, on September 6, 2025, the artists' group provided a press release to technology media outlet Ars Technica announcing details of the settlement. The press release stated, 'We have successfully recovered the largest publicly announced settlement in the history of copyright litigation in the United States ($1.5 billion).' The settlement covers '500,000 works of pirated content that Anthropic used to train its AI.' If the court approves the settlement, the artists' group will be able to receive $3,000 (approximately 440,000 yen) for each work that was plagiarized. According to the artists' group, 'the final amount per work may be even higher depending on the number of claims filed.'
Anthropic has already agreed to settlement terms, but the settlement requires court approval before it can become final. According to the press release, preliminary approval could come as early as the first week of September 2025, but a final decision could be delayed until 2026.
Justin Nelson, the attorney representing the three authors who filed the original lawsuit that sparked the class action, acknowledged that if the settlement, a 'first' of the 'AI era,' is approved, the amount paid 'would far exceed any other known copyright recovery.'
'This settlement provides meaningful compensation for each classwork and sets a precedent for requiring AI companies to pay copyright holders,' Nelson said. 'This settlement sends a strong message to both AI companies and creators that it is wrong to remove copyrighted works from pirate sites.'
Mary Rathenberger, CEO of the Authors Guild , said this is 'a great outcome for authors, publishers, and rights holders overall.' 'It shows the serious consequences of companies pirating authors' works to train AI, taking them away from those who can least afford it.'
Maria Pallante, president and CEO of the Association of American Publishers, agreed that the settlement was 'beneficial' to all involved beyond financial considerations, saying, 'The proposed settlement is of great value in sending a message that AI companies cannot illegally obtain content from shadow libraries and other pirated sources as building blocks for their models.'
A notable aspect of this lawsuit is that the settlement allows authors to retain their rights and legal claims in works that are not the subject of the lawsuit, and does not waive past or future claims in works that Anthropic may have infringed.
If a settlement is tentatively approved in the coming weeks, authors can search the following websites to see if their work is included in the class action and whether they will be eligible for payments. Authors seeking compensation can also provide their contact information to be notified when a settlement is reached.
Home | Bartz, et al. v. Anthropic PBC
https://www.anthropiccopyrightsettlement.com/
Ars Technica wrote, 'Anthropic is likely breathing a sigh of relief that it has avoided protracted litigation and further payments for pirated books. But the entire AI industry must be horrified by this settlement. Supporters have pointed out that it could set a precedent that could financially ruin AI startups like Anthropic.'
When Ars Technica reached out to Anthropic for comment, the company's deputy general counsel, Aparna Sridhar, said, 'Today's settlement, if approved, will resolve the remaining plaintiffs' claims. We remain committed to developing safe AI systems that enhance the capabilities of people and organizations, advance scientific discovery, and help solve complex problems.'
・Continued
Judge puts a hold on Anthropic's copyright infringement class action settlement agreement - GIGAZINE
Related Posts:
in Software, Posted by logu_ii