Художники судятся с Google, Meta и Anthropic за обучение ИИ на их книгах

Writers and

Artists Sue Google, Meta, and Anthropic Over AI Training on Their Books

In July 2026, writers and artists launched a wave of lawsuits against Google, Meta, and Anthropic, accusing the AI companies of training their models on pirated copies of their works. The trigger was a database published by The Atlantic, which lets authors search for their own works among the texts used for training.

What sparked the wave of lawsuits

The trigger was The Atlantic's publication: the outlet released a searchable database of works that ended up in AI models' training sets. Authors started typing in their names — and finding themselves. Documentary writer Kirk Wallace Johnson found both of his books there, "The Feather Thief" and "The Fishermen and the Dragon," nonfiction works that, he says, each took five to six years of research and work. In effect, his texts had been pirated and fed to a chatbot.

Why the authors went to court

The authors accuse Google, Meta, and Anthropic of using their works to train neural networks without permission or payment. For many writers, this concerns years of labor that ended up in the training data of commercial models.

• Defendants in the lawsuits — Google, Meta, and Anthropic
• Trigger for the public exposure — The Atlantic's searchable database of training texts
• Example plaintiff — documentarian Kirk Wallace Johnson, author of "The Feather Thief" and "The Fishermen and the Dragon"
• Author's investment — five to six years of research and writing per book
• According to The Verge, artists are already winning some of these lawsuits

Johnson describes his feelings as a "cocktail" of several emotions.

"Anger at the audacity of this theft, anxiety about what it means for writers, and a healthy desire for revenge against these giant corporations that have become fabulously wealthy," —

Kirk Wallace Johnson, writer, in a piece for The Verge.

What this means for AI companies

The wave of lawsuits is moving the debate over copyright and model training out of public discussion and into the courtroom. Previously, artists and writers could only voice outrage on social media; now they have a verification tool — The Atlantic's database — that makes it easy to prove a specific work ended up in the training data. As The Verge notes, authors are already winning individual cases, which means companies' claims of "fair use" are no longer a foolproof defense.

What it means

The conflict between authors and the AI industry is entering a stage where claims are backed by evidence and court rulings, not just statements. For Google, Meta, and Anthropic, this means growing legal and financial risks around what data their models were trained on.

• Meta is recognized as an extremist organization and is banned in Russia.