Meta He is immersed in a copyright rape trial to train his artificial intelligence Llama. Yesterday several emails have been made public that show that Meta It was downloaded by BitTorrent and similar tools complete databases of pirate bookshow Whisperwhich are offered on well -known piracy websites.
The accusation, an association of writers whose books have been used to train artificial goal intelligence Without their permission (and without paying them), he has presented these emails as proof that the goal was downloaded through torrent files dozens of millions of books, using techniques to hide their identity.
The severity of the evidence goes beyond Rirate booksbecause when using torrent files, to download you have to share, so can be accused of participating in the distribution of content protected by copyright.
The Pirate of Meta to train its AI
The revelation of which target had used torrent files for downloads Millions of pirated e-books He met last month, and the goal itself acknowledged that it was true. But it has been yesterday when, thanks to some emails sent to the Prosecutor’s Office, the exact data is known. You can consult those emails in this link of Ars Technica.
According to the document, «at least 81.7 TB of multiple libraries in the shadow were downloaded through the Anna’s Archive site, including at least 35.7 Terabytes de Data from Z-Library and Liben. Meta also previously torrenteed 80, 6 LIBGEN DATA TERABYTES «.
Goal was aware of what he did. In the emails you can read from employees who complain about hacking books, to Tips for downloading torrents from servers that were not Facebookto prevent them from tracking them through the seeds.
«Making torrents from a corporate laptop does not seem good to me,» writes with a sneer the Meta Nikolay Bashlykov research engineer to open 2023, using a smiling emoji. In the same message, he indicates that «I am concerned about the use of target IP addresses to download through torrents pirate content.»
Goal, therefore, was aware that I was popular booking, but the operation was completed until download dozens of millions of ebooks protected by copyright.
The authors’ strategy is why Goal not only pirate books, but distributed them while downloading them through Bittorrent or similar.
The goal defense, meanwhile, continues to insist that The download of the Libgen database was made within what it considers «fair use». A very peculiar way to call what it is to train artificial intelligence to earn money …
The defense has also declared that «the plaintiffs do not allege a single case in which any part of any book was, in fact, downloaded by a third party of the finish line through Torrent, much less than the books of the plaintiffs were some of some way distributed by goal «.
The authors may not demonstrate that, exactly their books were shared by goal, but the truth is that, because of the way the Torrent protocol works, To download you have to share what you are downloading.
A very murky matter that demonstrates that the big technological companies have no problem skipping law and copyright, to meet their objectives.
And although it is demonstrated that Meta downloaded dozens of millions of books protected by copyright of websites, through files Torrentsurely this trial will end like all: paying a millionaire compensation, and something else. Money is left over …
Know How we work in NoticiasVE.
Tags: Artificial intelligence, goal, viral, industry, ebooks, copyright