Sports and Competition Sports and Competition
Wed, February 12, 2025

Facebook employees used Torrent to download 81.7TB of data to train its LLaMA: Court papers


Published on 2025-02-12 05:45:29 - MSN
  Print publication without navigation

  • Meta, Facebook's parent company, is facing a class action lawsuit for copyright infringement and unfair competition, alleging the use of 81.7TB of pirated data to train their AI model. Despite concerns from researchers,

According to court documents, Meta Platforms Inc., the parent company of Facebook, utilized torrent technology to download approximately 81.7 terabytes of data for training its AI model, Llama. This data was sourced from various online repositories, including those containing copyrighted books, which has led to legal scrutiny. The documents reveal that Meta employees accessed this data through torrents, a method often used for large file transfers, to amass the necessary information for enhancing their AI capabilities. This practice has raised concerns about copyright infringement and the ethical implications of using potentially pirated content to train AI systems. The case highlights the ongoing tension between tech companies' data needs for AI development and intellectual property rights.

Read the Full MSN Article at:
[ https://www.msn.com/en-in/money/news/facebook-employees-used-torrent-to-download-81-7tb-of-data-to-train-its-llama-court-papers/ar-AA1ySUKT ]
Contributing Sources