tl;dr: A federal judge ruled in Bartz v. Anthropic that Anthropic illegally pirated books from shadow libraries like LibGen to train Claude, resulting in a $1.5 billion settlement covering ~500,000 works, though the judge found training on legally acquired books to be fair use. Stanford professors Tobias Wolff and Chang-rae Lee opted out and filed a separate suit, Cruz v. Anthropic, seeking up to $150,000 per infringed work versus the ~$3,000 per work in the class settlement. The Authors Guild is also lobbying Congress for legislation mandating AI training licenses.