Unsealed filings show Microsoft exec called AI training theft of labor
The quote comes from a separate news-publishers' case, while authors' own summary-judgment motion sits pending since September 4

Somebody sent you the quote already, didn't they? A Microsoft exec calling AI training "the largest theft of labor in human history." It's the kind of line that makes you want to forward it to every author you know with three exclamation points. Before you do — slow down and get the facts straight, because this one's easy to misread.
That quote, from Microsoft's Director of Applied Science, Brent Hecht, comes from newly unsealed filings in the news-publishers' lawsuit against OpenAI and Microsoft — the case brought by The New York Times and others over scraped content and bypassed paywalls. It is an allegation surfaced in court documents, not a finding by any judge. No court has adopted it as fact.
Here's the part that actually affects you as a book author: there's a separate strand of this same consolidated litigation, brought by the Authors Guild, arguing OpenAI used copies of copyrighted books — including material pulled from LibGen — to train its models. That motion for summary judgment was filed on September 4, and it's still pending. A February 2026 court order already established, as fact, that an OpenAI employee downloaded books from LibGen and that the resulting datasets, Books1 and Books2, were used to train GPT-3 and GPT-3.5. That part isn't in dispute. What's still contested is whether doing that counts as infringement or protected fair use.
Internal messages in the authors' filings reportedly call the practice "sketchy" — but again, that's the plaintiffs' characterization, not an admission from OpenAI or Microsoft.
Adding another layer: the Department of Justice filed a Statement of Interest in the news-publishers' case, backing arguments favorable to OpenAI and Microsoft's fair-use defense. That's a federal body signaling there's a legitimate legal argument on the companies' side — at least in the news-scraping dispute. It doesn't touch the book-authors' case at all.
So here's what to actually do this week: stop treating the Hecht quote as proof of anything for your own claim. If you're a book author with work potentially in the LibGen-derived datasets, your leverage is the pending September 4 summary-judgment motion, not a viral quote from a different lawsuit. Watch for a ruling on that motion — it's the next real signal, not the headlines built on unsealed exhibits. And when you talk about this case with other writers, separate the two strands out loud. Confusing them muddies your own case and hands the companies an easy "that's not even our lawsuit" rebuttal.
From the Coral Wire edition of Sat, Sep 26 2026 · Coral Index 62/100 (Tailwinds building)
Coverage — 3 outlets
Every outlet found reporting this story.
- OpenAI Accuses Plaintiffs’ Lawyers Of Paying For, Hiding, And Then Laundering Sketchy Key Evidence In AI Copyright Case | Techdirttechdirt.com
- Judge Refuses to be “Jerked Around” By Freeman and AttorneysPublishers Marketplace
- OpenAI Microsoft Copyright Lawsuit Unsealed Filings Largest Theft Laborkarmactive.com