Anthropic bought and shredded millions of books to train Claude – court rules fair use

Anthropic spent millions of dollars buying books published before the AI era, which were then cut apart, scanned and sent for recycling solely to provide high-quality training data for its Claude AI model. A US court ruled the practice falls within the "fair use" doctrine under American copyright law. The case has reignited debate over intellectual property rights in the age of artificial intelligence.

More on this topic

💻 Technology Anthropic settles for $1.5B over pirated books used to train Claude AI Hacker News · 7d ago 💻 Technology Anthropic's $1.5bn copyright settlement approved; authors get ~$3,000 per work Ars Technica · 8d ago 💻 Technology Anthropic's Claude AI cracked post-quantum cryptography that stumped human researchers for years Decrypt · 21h ago

More on this topic

💻 Technology Anthropic settles for $1.5B over pirated books used to train Claude AI Hacker News · 7d ago 💻 Technology Anthropic's $1.5bn copyright settlement approved; authors get ~$3,000 per work Ars Technica · 8d ago 💻 Technology Anthropic's Claude AI cracked post-quantum cryptography that stumped human researchers for years Decrypt · 21h ago