In AI copyright case, Zuckerberg turns to YouTube for his defense

Meta CEO Mark Zuckerberg's deposition in the AI copyright case Kadrey v. Meta reveals his defense of using a data set containing copyrighted e-books to train AI models. Zuckerberg compared Meta's situation to YouTube's efforts to remove pirated content, suggesting that a blanket ban on using such data sets might not be appropriate. He acknowledged the need for caution when dealing with copyrighted material but claimed limited knowledge of the specific data set, LibGen, used by Meta. The case is part of a broader legal battle between AI companies and copyright holders, with Meta accused of using pirated content from LibGen and Z-Library to train its Llama models.

The plaintiffs, including authors Sarah Silverman and Ta-Nehisi Coates, allege that Meta cross-referenced pirated books with those available for licensing to decide on potential agreements with publishers. The amended complaint claims that Meta used LibGen for training its Llama 3 and potentially Llama 4 models, and that Meta researchers attempted to conceal the use of copyrighted materials. The case highlights ongoing tensions over the "fair use" defense in AI training and the legal implications of using pirated content.

Key takeaways:

Mark Zuckerberg defended Meta's use of a data set containing copyrighted e-books to train AI models, comparing it to YouTube's handling of pirated content.
The deposition is part of the AI copyright case Kadrey v. Meta, where plaintiffs argue against the "fair use" defense claimed by AI companies.
Meta allegedly used the LibGen data set, known for containing pirated works, to train its Llama AI models, raising legal concerns within the company.
New allegations suggest Meta cross-referenced pirated books with licensed ones and used other sources like Z-Library for training its AI models.

In AI copyright case, Zuckerberg turns to YouTube for his defense | TechCrunch

Key takeaways:

Comments (0)

Newsletter