Judge approves the $1.5B Anthropic settlement over pirated training books

Sources: TechCrunch, Washington Post, HN discussion
Summary: US District Judge Araceli Martinez-Olguin approved the class-action settlement in Bartz v. Anthropic, under which Anthropic pays about 1.5 billion dollars, roughly 3,000 dollars per book, to authors and publishers whose pirated works were used to train the Claude models. The ruling covers more than 482,000 books, about 91 percent of which have been claimed by rights holders now due payment. Plaintiffs' counsel called it the largest known copyright recovery in history, and it is the first major settlement among dozens of pending AI copyright suits. The case was filed in 2024 by novelist Andrea Bartz and two other authors.
Why it matters: A concrete, court-approved price on training a frontier model with pirated text sets a precedent every lab and every team building on scraped data now has to reckon with.
Follow-up: Watch the claims-distribution mechanics, whether the per-work figure anchors other pending AI copyright cases, and any appeal.