300 authors skip class action, sue Meta solo over AI training

Here's a sentence you don't expect to write about litigation: authors are turning down the easy money.

On Monday, roughly 300 authors filed individual copyright lawsuits against Meta in the Northern District of California, choosing to sue on their own terms rather than join the class action wave that's already swept through the AI industry. Their lawyers — Freedman Normand Friedland LLP and Stris & Maher LLP — say close to 1,000 more authors are lining up to file similar suits in the coming months. Fiction writers, journalists, the whole range.

Why skip the class? Because the math looks different when you go solo. The claim is the same one that's been made against Meta and other AI firms for a while now: that the company pulled hundreds of thousands of e-books from shadow libraries — the pirate archives of the publishing world — and fed them into its large language models without asking anyone's permission. But instead of opting into a shared pot, each of these authors is suing individually, and each is seeking statutory damages that can run as high as $150,000 per infringing work.

Compare that to the numbers we've already seen. The Anthropic class action settled for $1.5 billion, which worked out to about $3,000 per author who opted in. Meanwhile, in a separate case, Kadrey v. Meta, a judge found that plaintiffs hadn't shown enough evidence that Meta's AI model was actually hurting the market for their books — a narrow but real setback for that line of argument. Put those two results side by side and you can see why a lawyer might tell a client: don't settle for the class rate when the statute allows for so much more, assuming you can prove your case.

Meta's defense here is the one you've heard from every AI company sued over training data: fair use, because the copying transforms the material into something new. That argument hasn't gone away, and it isn't going away for these new filings either.

What this means for you: if you're an author and you suspect your work ended up in a dataset that trained a major AI model, you now have two real paths in front of you, not one. Class actions offer a known, modest payout. Individual suits offer a bigger ceiling — and a bigger burden of proof.

This week, if you've had any inkling your books might be caught up in this, start keeping your own paper trail: publication dates, ISBNs, any sign your work showed up in a shadow library. Four new cases were just filed — Reno, Sheard, Winspear, and Older v. Meta Platforms — and that list is about to get a lot longer. Watch where your name might fit on it.