Anthropic reached a $1.5 billion class-action settlement with authors whose copyrighted books were used to train Claude, with individual authors receiving approximately $3,000 per work after legal fees. A federal judge affirmed that Anthropic's use of copyrighted material qualified as fair use under law, but found the company liable for illegally copying and storing millions of pirated books, creating financial exposure that incentivized settlement despite a favorable precedent on the underlying legal question. The settlement signals that while fair-use doctrine may shield AI training from copyright claims, practical liability for data sourcing and storage methods remains quantifiable and significant enough to compel seven-figure settlements. For financial institutions, this establishes that training-data provenance and chain-of-custody liability is now a measurable balance-sheet line item for frontier AI vendors, with precedent suggesting nine-figure settlement costs are plausible even when the core legal theory (fair use) prevails.
377 other tracked developments touching the same claims.