Why Authors Just Won $1.5 Billion

Some business news reads like a lullaby.

Morning Brew is the opposite.

A free daily newsletter that breaks down what’s happening in business and culture — clearly, quickly, and with enough personality to keep things interesting.

Each morning brings a sharp, easy-to-read rundown of what matters, why it matters, and what it means to you. Plus, there’s daily brain games everyone’s playing.

Business news, minus the snooze. Read by over 4 million people every morning.

Try Morning Brew for Free

Unsealed court documents reveal the wild lengths AI companies went to train their models — including physically destroying millions of books. But this chaos produced the largest copyright settlement in U.S. history and created a real framework for compensating creators. The AI you're using exists because of this messy process. Now we get to decide how it works going forward.

"Project Panama is our effort to destructively scan all the books in the world."

That's from an internal Anthropic document unsealed last week. The project involved buying millions of physical books, slicing off their spines with a hydraulic cutter, and scanning every page to train Claude.

One of those books? Ray Bradbury's Fahrenheit 451 — a novel about burning books.

You can't make this up. But here's what's interesting: this chaos just produced something genuinely useful — the first real framework for compensating creators whose work powers AI.

Anthropic downloaded 7 million pirated e-books, then launched Project Panama to buy and scan millions more. Pages were recycled afterward.

Meta torrented 81.7 terabytes of pirated books. Internal messages show employees debating the ethics in real time. One wrote: "I feel that using pirated material should be beyond our ethical threshold." They did it anyway.

OpenAI used similar pirate libraries early on but claims it deleted everything before ChatGPT launched.

It's messy. It's also how transformative technology has always worked. The printing press, photography, radio — all moved faster than the rules. Then the rules caught up.

That's happening now.

In September 2025, Anthropic settled for $1.5 billion — the largest copyright settlement in U.S. history.

It established a floor. Authors receive approximately $3,000 per work. Not retirement money, but a baseline future deals will build on.

It proved the legal path works. Three authors filed suit. It became a class action covering 500,000 works. Creators organized, courts took it seriously, and a major AI company wrote a very large check.

It drew a clear line. Training on legally acquired content = fair use. Training on pirated content = not. That gives creators real leverage going forward.

The AI you use right now — the one that helps you draft emails, brainstorm ideas, and learn new things — exists because of this messy process.

These models are useful precisely because they absorbed so much human knowledge. You benefit every time you use AI to summarize a dense report, get unstuck on a project, or automate tedious work.

The question isn't whether the technology should exist. It does. The question is: now that we're here, how do we make it work for everyone?

Licensing deals are multiplying. New tools help creators track and protect their work. The legal precedent is set. We're watching the rules get written — and for the first time, creators are in the room.

A free tool that lets creators search whether their work appears in AI training datasets.

Check if you're owed money. If you've ever published a book, visit anthropiccopyrightsettlement.com and search for your work. Takes two minutes — you might have $3,000 waiting.