Books by 21 Korean authors used to train Anthropic's Claude fall under $1.5 billion copyright settlement
2026-10-17 | USA | incident | economic
Our take
Twenty-one books by Korean authors went into Claude's training data without anyone asking. The authors now get to file a claim, which is the asking, done afterward.
The facts
US AI company Anthropic used 21 English-language books by Korean authors without permission to train its Claude chatbot, resulting in a US court ruling of copyright infringement. Anthropic agreed to pay $1.5 billion in damages, with affected Korean authors eligible for compensation. The incident highlights AI-driven intellectual property violations.