Anthropic Admits to Copying Books en masse for Claude—Can Fair Use Save It?

A U.S. federal court issued its first major order in the Claude AI copyright case. Authors allege Anthropic pirated or scanned their books for LLM training. The court ruled that while Claude’s use of scanned books for training was fair use due to its transformative nature, the creation of a central library using pirated books was not justified. A class certification motion remains pending.

Each work selected for training any given LLM was copied in four main ways — and in fact so many times that Anthropic admits it would be impractical even to estimate.

First, each work selected was copied from the central library to create a working copy for the training set.

Second, each work was cleaned to remove a small amount of lower-valued or repeating text (like headers, footers, or page numbers), with a “cleaned” copy resulting. If the same book ...