Meta deliberately used pirate books for AI education, Senate Hearing

Senate Josh Hawley, R-Mo., Called artificial intelligence More than 200 TERABITIES of META, the authors, the roles are alleged to have works published without a dime to make EU smarter

This Senate Justice Subcommittee Crime and anti -errorism About a court preparation to explore the adoption of copyrighted works for the EU industry for AI education.

Hearing was held in San Francisco a few weeks after two federal judges AI companies You can use books to bring up AI systems such as meta and anthropical.

How Bhamati Viswanathan Humanli described the professor and legal scientist during hearing on Wednesday Technological companies Get a great data set to bring up AI systems, everything is not added is the pirate work.

Amazon CEO says that AI will reduce the workforce

Companies do not receive books from the authors such as David Baldancci, who testified at the court hearing. Instead, theft or pirate, the authors of the licensed material were allegedly explained by the authors of the material.

Attempts were made to do the criminals, but Viswanathan said: “This is like a game of a mole – you get one of the ones you need to beat him again in the jurisdiction.”

Viswanathan said the criminal copyright liability had two bending. You have to do it in a prong, do you have to do it intentionally and do it for the second Prong, a commercial advantage or profit. In Meta caseHe noted that the company did it for a commercial advantage or profit.

Viswanathan claimed to know what the visible works deliberately saw, Meta knew what it did.

Amazon announces $ 20B investment in Pennsylvania for AI Data Centers

“Even (Meta CEO) Mark Zuckerberg (Hey, ‘Hey, Hey?’ And he did not know what he did not know.

Hawley, then Maxwell Pritt, who represents several authors about the theft of copyrighted works, began to ask Maxwell Pritt.

Pritt claimed that many of the most important “public criminal enterprises” called Meta’s “shadow libraries” were torrented by more than 200 terabytes of copyright.

The lawyer said that Meta billions of work and the authors did not do anything to the book. When Meta’s authors investigate their payment, Pritt said, “No”.

Trainings on the authors of the ruling AI companies, unauthorized books

“Early, they investigated the license. They set the part-time time of two individuals to try to initiate the license and when it will be very long and turns it into this piracy,” he said. “At that time, there were social documents, which was intended for licensing tens of millions of millions, in this period, and there were ten million people.”

During the court, Hawley showed pictures of text messages related to the adoption of works published among meta workers and EU engineers.

Artificial Intelligence requires an electrical network update

A meta engineer working on the AI project, “I don’t think we should use pirate material. I really need to take a line there.” They went to say that he felt that the pirate was using the material, overturned the team’s ethical limit.

In the same conversation, another person answered: “This is the piracy of the issue,” he said.

And another, “We want to get a book and be beautiful here … But you need to do it and make some corners to let bad people win.”

Pritt, “bad children” said there were other AI rivals.

“Yes, this, of course, is one of many documents that show that they are pirated sites with copyrighted materials, and they received them for free.

Hawley shared additional messages between meta workers.

“Torrent I’m not sure that the methane will use IPS to load with pirate content,” Ahah “wrote.” Another answer: “I am interested in starting to look at some examples, but I feel allowed to start looking.”

“Ahah, Yes, I think the sealing from corporate laptops does not feel right.”

In another line, employees said they could not use Facebook servers, because the loader’s back to Facebook.

Artificial intelligence, large technical partnership with nuclear energy manufacturers

“We have Meta workers who said they know they are, they think, think of their ethical behavior, they think it is illegal and avoid trying to create a paper path.” “They are trying to hide it. It doesn’t sound like a fair use of me.”

Professor Edward Lee, in the case of US meta platforms, the United States agreed with the approach to the district of the district Vince Chabria, he said.

Xhabria told the authors in the last month in the government that the meter did not provide enough evidence to dilute the market for sufficient work for a copyright violation.

“This decision said,” he said, “said, the use of copyrighted materials to cultivate the meter’s language models.” Reuters. “It stands for suggestions that these bidders do not have to develop in the wrong arguments and support the right support.”

Lee, the distribution case is still alive and the sliking aspect violates and not used fairly.

“I’ll say this. If this is not a violation, the Congress needs to do something,” Hawley said. “Almost, if the world’s largest corporation is the largest corporation worth the trillions, can the individual author take the work of the individual? to change the law“

Electric prices are higher for American houses: Driving costs are higher here

In addition to the benefit of the meter, along with the administration of the anthropy, the sentence came to the end of last month. The US District Judge William Alsup, writers referred to the “fair use” of books to cultivate a large language model of Kirze by Andrea Bartz, Charles Graver and Kirk Wallace Johnson.

But the copying and storage of Alsup, anthropy, “Central Library” Copying and storage of more than 7 million pirate books in the “Central Library” violated the author’s copyright and was not just used. The judge ordered how much an anthropic debt was to determine for violating a trial in December.

Fair use is the main legal protection for technological companies and the decision of Alsup is the first address in the context of a generative AI.

AI claims that they can use the AI industry, which are forced to use copyrighted material to create new, transformative content of systems and to create copyright owners for their work.

Ticker	Safety	Last	Change	Change%
Meta	Meta Platforms Inc. Inc.	702.91	-7.48	-1.05%

Including anthropic and other prominent AI companies Openai and Meta platformsMillions of books are accused of downloading pirate digital copies to train their systems.

The US Copyright Law says the reluctant copyright infringement can be based on legal damage to $ 150,000 for each job.

Click here to read more in the fox work

Copyright owners illegally copy their work to create a competitive content that threatens the livelihoods of AI companies.

Others, like the White House, as AI Czar David sacks, the United States can lose the US race without it without it.

“It is very important to end with a definition of a sensitive fair use, as in this anthropic situation, because otherwise we will lose AI race for China,” he said.

Fox Business’ Pilar Arias has contributed to this report.