Meta Admits Torrenting Pirated Books, Says It Did Not ‘Seed’ Them

Meta has filed a court document admitting it pirated terabytes worth of books, but defending itself on the premise that it did not “seed” the books.

“Seeding” is a term that refers to making making files available for download via torrent, a decentralized method of downloading and sharing files. In its filing, via Ars Technica, Meta freely admits it download a gargantuan quantify of books, but says the fact that it took measures to avoid seeding them means it does not hold any legal liability.

Meta’s Motion established that Plaintiffs lack standing and cannot state a claim for violation of Section 1202(b)(1) of the DMCA or the CDAFA. Focused on “torrenting,” which is a widely used protocol to download large files, Plaintiffs push a narrative that ignores evidence in their possession, including a detailed expert report, showing that Meta took precautions not to “seed” any downloaded files. Plaintiffs, thus, plead no facts to show that Meta seeded Plaintiffs’ books – a claim that Meta will address at summary judgment. For now, the Court should dismiss Plaintiffs’ unprecedented and untenable application of both Section 1202(b)(1) and the CDAFA to the allegations in the TAC.

Meta goes on to say that removal of information, such as copyright management information, doesn’t constitute a violation of the DMCA.

Mere removal of redundant information from training data, including copyright management information (“CMI”) in the form of boilerplate legal disclaimers or the copyright symbol, is not a violation of Section 1202(b)(1) of the DMCA. To survive dismissal, Plaintiffs were required to plausibly plead facts showing that Meta removed CMI with knowledge or reason to know that it would “induce, enable, facilitate, or conceal” infringement of Plaintiffs’ copyrighted works. They still cannot so – despite exhaustive discovery – because their allegations don’t add up. They cannot show that CMI removal enabled or concealed infringement by third parties (no such infringement is even alleged). They also do not allege any facts showing that CMI removal enabled or concealed any infringement by Meta, whose public disclosure of Llama’s training data prompted this copyright infringement suit. Separately, their Opposition confirms that Plaintiffs cannot allege a concrete or particularized injury caused by the supposed violation of Section 1202(b)(1), warranting dismissal with prejudice for lack of standing.

Meta also challenges the plaintiffs’ claims that they suffered “concrete” injury as a result of the company pirating their books has no merit. The entire document is a fascinating read, one in which Meta carefully avoids any variation of the words “pirate” or “pirating.”

Ultimately, Meta’s case will likely have a profound impact on the AI market. The company’s AI models are already open-source and freely available for others to use. At the same time, there’s no denying that Meta pirated millions of books to train its AI models.

If the court rules in favor of the plaintiffs, it will be a devastating blow to the entire AI industry, especially those companies that charge for their AI models, since—at least on the surface—unjust enrichment could become a much bigger factor.

from WebProNews https://ift.tt/QKi3VAO