# Publishers and authors say Meta knew training its AI on their books and articles would be a copyright infringement, but did it anyway.

*genai · news · 2026-05-05 · AM New York*

## Key points

- Meta allegedly masked its IP and removed copyright notices when scraping pirated sites for Llama AI training.
- Internal Meta communications show employees agreed not to disclose use of LibGen and other pirated sources.
- The lawsuit claims Mark Zuckerberg personally halted licensing negotiations to pursue pirating publishers' works.
- Llama AI is estimated to generate $460 billion to $1.4 trillion in profit by 2035.

**Companies:** Meta, McGraw-Hill, MacMillan Publishing, Hachette Book Group
**Countries:** United States

[Read the full story on AM New York](https://www.amny.com/news/publishers-allege-meta-ai-stole-books/)

---

Canonical: https://newsio.io/n/cd054c4f-7366-478b-861e-9db9d4822cd5/publishers-and-authors-say-meta-knew-training-its-ai-on-their-books-and-articles
Summarized by Newsio from AM New York. https://newsio.io/how-it-works