Anthropic Wins Round One in AI Book Battle

In a major win for AI developers, a federal judge just ruled that it’s totally legal for AI models to learn from books even without permission as long as they’re learning, not copying. That means Anthropic’s AI is off the hook for reading pirated novels to train its writing skills. Think of it like a robot book club, not a copyright thief. But there’s a plot twist: while teaching an AI to write like a human is fine, saving 7 million full pirated books in one giant library? Not cool. That, the judge says, crosses the copyright line.

Key takeaway:

• Training = fair use

• Pirate hoarding = infringement.

This is the first major ruling on AI and copyright—and it’s setting the rules for everyone else. Next up: the piracy trial kicks off in December, with potential damages of up to $150,000 per book. Yes, per book. Stay tuned, this courtroom drama is just getting started.