Anthropic moves to dismiss authors' copyright suit on timing

Anthropic just told a federal court the real problem with the authors' lawsuit isn't what the company did — it's when the writers noticed.

In a motion filed Wednesday, Anthropic asked the court to throw out copyright infringement claims from a group of authors and publishers, arguing the case was filed too late. The company's logic is almost jiu-jitsu: it's using the plaintiffs' own complaint against them. Anthropic says that complaint cites public disclosures about its training practices from years before the suit was filed — which means, in Anthropic's telling, the writers can't now claim they were kept in the dark. "Plaintiffs can't claim they waited so long due to the concealment of Anthropic's training data since their own pleading cites public disclosures that made those facts available years before the suit," the company said in its filing.

Here's why that matters more than it sounds. This is a statute-of-limitations argument, not a fair-use argument. Anthropic isn't (yet) asking a judge to rule that training on copyrighted books without a license is legal. It's asking the judge to rule that the writers waited too long to complain — full stop, case closed, no verdict on the underlying question ever gets rendered. That's a pattern showing up across the AI copyright docket right now: procedural challenges first, substance later, if ever.

For you as a working author, the lesson isn't really about this one case — it's about what you do the next time you learn something new about how an AI company used your work. The moment public information about a company's training practices surfaces — a report, a disclosure, a leaked dataset list — that clock may start ticking whether you noticed it or not. Waiting to act because you want to "see how things play out" is exactly the posture Anthropic is now using against these plaintiffs.

So here's what to actually do this week: if you're part of any AI-copyright litigation or class action, or thinking about joining one, don't sit on news about a company's training data. Flag it to your Guild rep or your attorney the day you see it, and get it documented. And if you've been putting off finding out whether your work is implicated in any of these training sets, stop putting it off — a delay that feels reasonable to you can look, in a courtroom, like a missed deadline.

Watch for the ruling on this motion. If Anthropic wins on timing, it tells every other AI defendant that procedure beats principle — and it tells every author that the clock on your claim may already be running.