A federal judge granted final approval this week to a $1.5 billion settlement in a class-action lawsuit against Anthropic.
A class of authors allege the AI company illegally downloaded their copyrighted books from libraries LibGen and PiLiMi, which the company then used to build its book collection, Mashable reported.
The San Francisco court called it the largest copyright class action settlement in history.
The crux of the complaint was how Anthropic acquired the copyrighted material and not whether training its AI models with that material is lawful. An earlier ruling cited throughout the order had determined that the training itself constituted fair use.
The settlement will allow the authors and publishers whose pirated works ended up in Anthropic's library to claim approximately $3,000 per book, which is about four times the typical minimum for copyright infringement cases. Over 440,000 books have already been claimed.
The settlement also requires Anthropic to delete the material it downloaded.