March 27, 2011 - Technology media outlet WinBuzzer published a blog post yesterday (March 26) reporting that new research has shown thatMeta Participation in the digital book piracy chain is far greater than expected.
In addition to the news that Meta is using pirated books from "shadow libraries" such as LibGen and Z-Library to train AI models, Meta is also using pirated books from "shadow libraries" such as LibGen and Z-Library to train AI models.About 30% of downloaded books were also re-uploaded via BitTorrent, objectively extending the distribution cycle of pirated works.
Experts say that although BitTorrent uploads data in slices by default, Meta's uploads are unusually high, and it's suspected of being a piracy network node in disguise.
As 1AI reported in February, Meta claimed in court documents that even though it downloaded 82 terabytes of pirated copyrighted material from its "shadow library" to train its LLaMA AI model, company employees took steps to...Ensure that no files are "seeded" during the download process..
But new research points to secondary uploading behavior that could turn this argument on its headFormer Australian Prime Minister Malcolm Turnbull is "extremely disturbed" to discover that his writings have been misappropriated, Pulitzer Prize-winning journalist Michael Chabon and others have filed a lawsuit, and a French publisher is accusing Meta of "massive plunder of copyrighted works. French publishers have accused Meta of "massive plundering of copyrighted works".
The European Commission has taken an interest in the matter and may impose severe penalties under the EU AI Act. If the court ultimately decides that Meta needs to obtain a formal license for its training data, it could reshape the rules of data access in the AI industry. It's not yet clear where the case will go, but the copyright dispute is destined to become a benchmark for intellectual property protection in the digital age.
