Anthropic to Pay $1.5B in Landmark Copyright Settlement

Anthropic has reached a $1.5 billion settlement with authors and publishers after it was revealed that its Claude AI was trained on pirated books taken from online shadow libraries such as LibGen and Books3. The agreement covers around 500,000 works, with authors expected to receive an average of $3,000 per book or story. This marks the largest copyright settlement in U.S. history. As part of the deal, Anthropic is required to delete all illegally obtained copies, certify their removal, and guarantee that future training data will come only from legally licensed sources. The settlement still requires court approval, and the final payout may increase if more authors join the claim before the hearing.

The case has sent shockwaves through the AI industry, establishing a strong precedent that scraping copyrighted works from pirate websites constitutes infringement, regardless of debates around fair use. Legal experts note that the outcome will likely push AI companies toward more transparent licensing models and legitimate datasets, while also encouraging creators to assert their rights more actively. For authors, this agreement is seen as a major victory in the ongoing battle against digital piracy, demonstrating that their work can and will be protected in the face of unauthorized exploitation.