Anthropic agreed to a record $1.5 billion settlement for illegally downloading millions of pirated books to train its A.I. models. Though courts affirmed that training on legally obtained books is fair use, the ruling signals major liability for using shadow libraries. The case may push A.I. firms toward licensing and greater copyright compliance.

The article explains that most countries—including the U.S., EU, and UK—do not grant copyright to purely AI-generated works, as they lack human authorship. However, China and Japan may allow protection if there is clear human creative input. As AI tools evolve, laws may need to adapt to recognize shared human–AI creativity.