Meta Sued by About 300 Authors Over AI Training

Meta has been hit with a new wave of copyright lawsuits from authors who allege it used pirated copies of their books to train its AI models. About 300 authors have signed on to pursue individual claims, Bloomberg Law reported, citing their lawyers. The complaints were filed on Monday 19 October in the US District Court for the Northern District of California.

The cases mark a change of tactics. Most authors who have sued AI companies so far have grouped together in class actions, but these writers are bringing their own claims instead. For Meta and the rest of the industry, that could mean defending many separate cases, each with its own potential damages, rather than one large one.

In brief

About 300 authors are pursuing individual copyright claims against Meta over AI training, according to their lawyers, as reported by Bloomberg Law.
They allege Meta downloaded e-books from “shadow libraries” to train its models; these are allegations that have not been tested in court.
Meta did not immediately comment, and the lawyers plan claims for close to 1,000 more authors in the coming months.

What the authors allege

According to Bloomberg Law, “around 300 authors signed on to pursue individual claims over allegations the tech giant took copies of their books and used them to train its large language models without permission”. They are represented by the law firms Freedman Normand Friedland and Stris & Maher. The authors range from novelists to journalists, Bloomberg Law said.

The core claim mirrors earlier cases. The suits allege that Meta “downloaded hundreds of thousands of e-books from illegal online databases known as shadow libraries” and used them to train its AI models, Bloomberg Law reported. None of this has been proven in court.

Court records confirm that the filings exist. CourtListener’s docket for Winspear v. Meta Platforms, case 3:26-cv-12100, shows a complaint filed on 19 October by 60 named plaintiffs, including Jacqueline Winspear, under the copyright infringement statute, with a jury demand. A separate docket, Reno v. Meta Platforms (3:26-cv-12131), was filed the same day. Bloomberg Law also lists Sheard v. Meta Platforms and Older v. Meta Platforms among the new cases.

What is at stake in damages

The authors are seeking statutory damages, according to Bloomberg Law. These can amount to a maximum of $150,000 per infringing work. That figure is the legal ceiling a court can award for each work, not a sum each author is guaranteed or has necessarily claimed.

That is where the individual approach matters. In a class action, many authors share one outcome. Bloomberg Law noted that the first authors’ class action against Anthropic ended in a $1.5 billion settlement, which would yield roughly $3,000 to each author who opted in. By suing individually, these authors keep their own claims and their own route to damages.

The lawyers plan to go further. The two firms said they intend to file individual lawsuits against Meta on behalf of close to 1,000 more authors in the coming months, according to Bloomberg Law. Meta “didn’t immediately return a request for comment”, the outlet said.

Fair use remains the central fight

The likely battleground is fair use. According to Bloomberg Law, AI companies including Meta have argued that training falls under copyright law’s fair use doctrine because they transform the copies into an entirely new technology.

The courts have so far given mixed answers. In an earlier proposed class action against Meta, Kadrey v. Meta, the judge found that the plaintiffs had failed to provide evidence that Meta’s AI model was harming the market for their works, Bloomberg Law reported. TSN has also covered a Third Circuit ruling that training an AI on Westlaw headnotes was not fair use, in a case involving a different kind of AI tool.

The pressure on how AI companies source training data is coming from several directions. TSN has reported on Meta removing a “train” sentence from its new commercial terms after a backlash, and on Google negotiating to buy Korean books for AI training.

Our read

This is TSN’s analysis. Individual claims take away the option of resolving everything through a single class settlement. If hundreds or eventually around 1,300 authors each press their own claims, Meta faces a long tail of litigation even if it wins on fair use in some cases. The next markers are Meta’s first responses in court and whether the promised further filings arrive.