A federal judge sides with Anthropic in lawsuit over training AI on books without authors' permission

“That Anthropic later bought a copy of a book it earlier stole off the internet will not absolve it of liability for theft but it may affect the extent of statutory damages.”

Judge Alsup ruled that training on books is fair use. He also ruled that keeping millions of pirated books from LibGen and other shadow libraries is not, and that part goes to trial. The company bought and scanned print books later, but the judge says that does not undo the piracy. Training won, and the library it was built from is on the hook.