Суд утвердил рекордное соглашение Anthropic: $1,5 млрд авторам книг за пиратские копии в датасетах

Federal court in the US has finally approved Anthropic's $1.5 billion settlement with book authors — TechCrunch reported this on July 20, 2026. This is the largest payout in a copyright case in US history and the first settlement of this scale between an AI company and rights holders.

What Anthropic is paying $1.5 billion for

Anthropic is paying for downloading pirated copies of books from shadow libraries and storing them in an internal dataset for training Claude models. The class action Bartz v. Anthropic was filed in 2024 by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson: they proved that the company downloaded millions of books from sources like LibGen without authors' permission.

• The settlement amount is $1.5 billion, roughly $3,000 per work
• The compensation covers about 500,000 books
• Judge William Alsup gave preliminary approval in September 2025
• Final approval took place in July 2026
• Anthropic also committed to destroying the pirated copies from its storage

Payments will be distributed through a special fund: authors and publishers have been filing claims since fall 2025. Some rights holders objected to the terms — in their view, $3,000 doesn't cover the value of a book to a commercial model — but the court found the settlement fair and reasonable.

Why the AI training dispute isn't closed

The settlement closes a specific case but doesn't answer the industry's main question — whether it's legal to train models on copyrighted texts. Back in June 2025, Judge Alsup ruled that training on legally purchased books is fair use, while downloading pirated copies is an infringement. The $1.5 billion payout relates specifically to piracy: had the case gone to trial, statutory damages threatened to reach hundreds of billions of dollars — US law allows up to $150,000 for each willful infringement.

"The final approval closes one case but doesn't resolve the broader issue of using copyrighted works to train AI models,"

TechCrunch's report states.

What other AI companies face

Other AI developers face dozens of similar lawsuits: according to industry estimates, more than 40 copyright cases are underway in US courts against AI companies, including OpenAI and Meta*. The most notable is The New York Times' lawsuit against OpenAI and Microsoft, filed in December 2023 and still unresolved. In a parallel case against Meta, Judge Vince Chhabria sided with the company on the fair use question in June 2025, but stressed that the ruling applies only to the specific plaintiffs. Now rights holders have an argument: even the wealthiest AI labs are willing to pay billions rather than let a case about pirated datasets go to a jury.

What this means

The era of free data for AI training is ending: training on legally obtained copies is protected as fair use, but pirated datasets will cost billions. The market is already restructuring — publishers and media are signing licensing deals with OpenAI and Google, and the price of legal training data is rising.

Frequently asked questions

How much will each author receive?

About $3,000 per book before deducting legal costs: the $1.5 billion fund will be distributed among rights holders of about 500,000 works that ended up in Anthropic's pirated dataset.

Did the court rule that training AI on books is illegal?

No. Judge William Alsup ruled back in June 2025 that training models on legally acquired books is fair use. Anthropic is paying for downloading and storing pirated copies, not for the act of training itself.

*Meta is recognized as an extremist organization and is banned in Russia.