AI Piracy Settlement
Judge Approves Anthropic's Record $1.5 Billion Settlement in AI Copyright Case
Anonymous
Key Points
1
The attorneys’ fees and litigation costs from the settlement total about $122 million, with roughly $3,100 allocated per author for more than 447,000 works.
2
Anthropic is backed by major tech giants Amazon and Alphabet, a detail highlighted by coverage of the settlement.
3
Final approval was granted by U.S. District Judge Araceli Martinez-Olguín in San Francisco on July 20, 2026.
4
The settlement covers claims on more than 92% of the works—roughly 447,000+ of the 480,000+ works included—according to authors’ counsel.
5
The pirated copies were drawn from two shadow libraries—Library Genesis and Pirate Library Mirror—with Anthropic required to destroy the original files.
A federal judge approved a $1.5 billion settlement between Anthropic and a class of authors on July 20, 2026 — the largest copyright payout in U.S. history, according to Illawarra Mercury. The deal resolves claims that Anthropic used pirated books to train its Claude AI without permission.
U.S. District Judge Araceli Martinez-Olguín granted final approval in San Francisco. The settlement covers more than 480,000 works and requires Anthropic to destroy the pirated files it used in training, according to Bendigo Advertiser.
Pirated Libraries at the Heart of the Case
Anthropic pulled the books from two shadow libraries — Library Genesis and Pirate Library Mirror. Together, they held more than seven million pirated copies. A prior judge, William Alsup, found that training on the books was fair use. But he ruled that storing millions of pirated copies in a central library crossed the line, according to Singleton Argus.
Alsup has since retired. Judge Martinez-Olguín took over and granted final approval. Anthropic does not admit wrongdoing under the terms of the deal. The company is backed by Amazon and Alphabet, two of the world's largest tech firms, according to Eden Magnet.
Authors Get $3,100 Per Work on Average
The $1.5 billion fund covers more than 447,000 works — about 92% of all works included in the case. Each author gets roughly $3,100 per work on average. Lawyers and court costs will take about $122 million off the top, according to Muswellbrook Chronicle.
Some authors say the amount is too small. Others argue the deal favors lawyers over writers. But authors' counsel say the 92% claims rate shows strong support for the settlement among the class.
What the Settlement Does — and Does Not — Cover
The deal is narrow in scope. It only covers how Anthropic got the books — not how it uses the content going forward. Authors still have the right to sue over Claude's outputs or any works not listed in the settlement, according to Illawarra Mercury.
That means key legal questions remain open. Licensing ongoing training data is not addressed. What AI models can legally generate from copyrighted text is still unsettled. Lawyers say the case will shape how AI companies weigh legal risk around training data going forward.
A Landmark Case in the AI Copyright Era
The Anthropic settlement is one of many high-profile copyright lawsuits hitting AI firms right now. Writers, musicians, and news publishers have all filed suits against AI developers over training data. This case is the first to reach a final settlement at this scale, according to Bendigo Advertiser.
The ruling draws a clear line. Downloading pirated content to build a training library is a legal risk — even if training on it may be fair use. AI developers are now expected to review how they source their data in light of this outcome.
Sentiment
Mostly NegativeNEGATIVENEUTRALPOSITIVE
Publishers
17
Articles
145
Reach
162
A
AnonymousAuthor