Court Documents Reveal Meta Used Pirated Books to Train its AI Systems

Not a great week for Zuckerberg’s PR team, with a new report suggesting that Meta maybe he was using torrent software for illegally downloading copyrighted books, which he then added to the datasets used to train his AI models. Zuck himself may have approved such usage.

Probably not a great way to increase its appeal to developers in its apps.

According to a new report from Wiredthe discovery was included as part of a copyright case brought by a group of authors against Meta over the development of its AI datasets. This is shown by the documents released by the court Meta used a tool called “Library Genesis” (LibGen) to access pirated versions of books to help build its datasets.

According to Wired:

“These recently unredacted documents reveal exchanges between Meta employees uncovered in the discovery process, such as a Meta engineer telling a colleague that they were reluctant to access LibGen data because they were “torrenting with [Meta-owned] a corporate laptop is no good.”. They also allege that internal discussions about the use of LibGen data were escalated to Meta CEO Mark Zuckerberg (referred to as “MZ” in a memo filed during the disclosure) and that Meta’s AI team was “approved to use” the pirated material.

Yeah, it’s not looking great for Zuck, who is already facing significant backlash for his recent knee-jerk reaction to President-elect Donald Trump.

Meta announced last week a major overhaul of its content moderation processwhile also eliminating its fact-checking program, in favor of an X-style community notes system. Right-wing politicians, including Trump, have long criticized Meta’s approaches on both fronts, and the changes appear to be in line with what Trump may have asked of Zuckerberg when they met at the end of last yearshortly after winning the elections.

At one stage, Trump is threatened Zuckerberg with life imprisonment for what he saw as election meddling, following the suspension of Trump’s Facebook account in 2021 (amid unrest in the Capitol).

But now Zuckerberg is apparently regularly traveling to Mar-a-LagoTrump’s home base in Florida, while Zuckerberg also intends to have a front row seat at Trump’s inauguration ceremony next week.

The change in approach quickly changed the perception of Zuckerberg, who has spent years reshaping his public persona after past controversies. Indeed, Meta changed its company name to distance itself from scandals such as Cambridge Analytica, in an effort to avoid association with previous data and access issues.

And Zuckerberg has recently proven to be a more approachable human presence. But the past two weeks have seemingly revealed once again that Zuckerberg values ​​business success above all else and that his moral views are dictated by profit.

This latest revelation will further underscore that, showing that Zuckerberg is potentially willing to undermine artists for his own business gain.

Which, I guess, is not a big discovery in the end. CEOs are responsible for the success of the business and are beholden to their shareholders, so it’s no big surprise that they will act in the best interests of the company. But lately there’s also been more pressure to support more ethical companies, and findings like these, if proven correct, could have an impact.

But on the other hand, the media cycle is much shorter than it used to be, and controversies usually only last a day or so, until the next edition of the day takes over. So maybe it’s not as big a risk as it seems, and maybe Meta sees the possible development of more advanced artificial intelligence as more valuable than the risk of accessing a single data set.

After all, Meta has previously argued that using any publicly available material to train its AI tools is covered by “fair use,” a copyright clause that allows such use in certain cases, such as news reporting.

That seems like a pretty outrageous misinterpretation of that clause, but then again, Meta’s legal team has been pretty valiant in their attempts to justify the company’s use of massive data sets.

Brash, and in some respects arrogant, knowing that it has the resources to fight legal threats into oblivion and that the laws are not currently structured to cover uses like artificial intelligence training.

That bullying approach is perhaps more true to Meta and Zuckerberg’s actual style, which again undermines the more relaxed persona he’s cultivated in recent years.

After all, Zuckerberg is a capitalist, and in that sense, all we see is a businessman maximizing his opportunities. But this could at some stage cause a significant reaction against Meta’s products.