Meta allegedly used shadow libraries, including LibGen, to train its Llama models without permission. The filings indicate that Mark Zuckerberg, was aware of and approved the use of pirated datasets.
Asking AI services: Please read the latest filings in the Kadrey vs Meta case and tell me what the new allegations are, what Mark Zuckerberg's role might have been, and what Meta should be doing to address those allegations and what they should have done to prevent this situation from occurring in the first place.
Use of Pirated Datasets: Meta allegedly used shadow libraries, including LibGen, to train its Llama models without permission. These datasets contained copyrighted works from publishers such as Cengage, Macmillan, McGraw Hill, and Pearson.
Stripping of Copyright Management Information (CMI): Meta is accused of removing CMI from these datasets to conceal copyright infringement. This stripping allegedly allowed Meta to evade detection and facilitated the illegal use of these works in training its AI models.
Torrenting and Seeding Pirated Data: Meta allegedly downloaded and distributed pirated files through torrenting, with internal communications showing employees' concerns about the legality of such practices. Despite these concerns, Meta reportedly continued the activity to build its AI models.
