Ddos 
AI developer Anthropic has agreed to pay $1.5 billion to settle lawsuits brought by authors and publishers who accused the company of using millions of copyrighted books to train its Claude model.
Specifically, Anthropic allegedly downloaded material from free pirated book-sharing websites and incorporated this content into model training without authorization. Notably, senior executives were reportedly aware that pirated works were being used, yet still permitted employees to proceed.
Under the settlement, Anthropic will pay a total of $1.5 billion in compensation to affected authors and publishers — the largest copyright-related payout in U.S. history. On average, this equates to roughly $3,000 per book, though the actual figure per author will be significantly lower after publishers take their share.
It is worth noting that the court had previously ruled that Anthropic’s use of such data for model training constituted “fair use.” The company maintains this position but agreed to compensation to resolve outstanding claims.
The $1.5 billion will be paid in four installments:
- First installment: $300 million within five days of the court’s preliminary approval of the settlement.
- Second installment: $300 million within five days of the court’s final approval.
- Third installment: $450 million (including interest) within twelve months of preliminary approval.
- Fourth installment: $450 million within one year of preliminary approval.
Meanwhile, other copyright-related cases remain pending. For example, Reddit has accused Anthropic of scraping its posts without authorization for training purposes — unlike Google and OpenAI, which have reached licensing agreements with the platform.
Finally, while this settlement resolves the copyright dispute with authors and publishers, a separate case will examine allegations that Anthropic executives knowingly allowed the use of pirated materials. That trial is scheduled for December 2025.
Donate