The judge approves $ 1.5 billion in copyright regulations between the company AI Anthropic and the authors

San Francisco – A federal judge approved on Thursday $ 1.5 billion regulations between the artificial intelligence company Anthropic and the authors who allege that almost half a million books had been illegally hacked to form chatbots.

US District Judge William Alsup published preliminary approval on Thursday before the San Francisco Federal Court after the two parties worked to respond to his concerns concerning the regulations, which pays the authors and publishers about $ 3,000 for each of the books covered by the agreement. It does not apply to future work.

“This is a fair settlement,” said Alsup, although he added that distribution to all parties will be “complicated”. According to Justin Nelson, lawyer for the authors.

“We have some of the best lawyers in America in this courtroom and if someone can do it, you can,” said Alsup.

The association of American publishers described the colony “not major in the right direction in the holding of the developers of the AI ​​responsible for an imprudent and shameless violation”.

“Anthropic is hardly a special case in terms of violation. All the other major AI developers have formed their models on the backs of authors and publishers, and many have obtained these works from the most notorious counterfeit sites in the world, “said Maria A. Pallante, president and CEO of the group of publishers.

Anthropic, based in San Francisco, said that it was satisfied with preliminary approval.

“The decision will allow us to focus on the development of safe AI systems which help people and organizations to extend their capacities, to advance scientific discovery and to solve complex problems. As we have always maintained, the benchmarks of June of the Court in June that the formation of AI constitutes an equitable use transformer remains intact. Obtained, “said Aparna Sridhar, deputy for the anthropogenic deputy councils.

The guild of the authors, on the other hand, said that the regulations “marks an important step in the fighting of the authors against the theft of their work of the companies of the AI. It sends a clear signal to AI companies that the violation of the rights of the authors has a steep price and will undoubtedly push the companies of the AI ​​to acquire the books they wish legally, thanks to the license. ”

A file on Monday sought to convince the judge that the parties created a system designed to obtain a solid opinion from all the authors and publishers covered by the agreement, ensuring that they obtain their pot cup if they want to sign the regulation or opt to protect their legal rights in the future.

They also tried to ensure that the author’s group and publishers who break the agreement together does no “rear” transaction that would injure the less known authors.

The main concern of Alsup has been focused on how the complaint process will be managed in order to ensure that all eligibles know so that the authors do not “get the tree”. He had set a deadline on September 22 for submitting a complaint form to be re -examined before Thursday’s hearing to examine the settlement again.

The judge had raised concerns of two major groups linked to the case – the guild of the authors and the association of American publishers – working “behind the scenes” in a way that could put pressure on certain authors to accept the regulations without fully understanding it.

The lawyers of the authors said that in the file on Monday, they thought that the regulations would result in a high rate of complaints, respects existing contracts and is “in accordance with the regular procedure” and the directives of the court.

Alsup had dealt with the case a mixed decision in June, noting that the training of IA chatbots on books protected by copyright was not illegal but that Anthropic has acquired millions of pounds wrongly via pirate websites to help improve your Chatbot Claude.

The novelist with successful thriller Andrea Bartz, who continued Anthropic with two other authors last year, said in a declaration of justice before the hearing that she strongly supports the settlement and will work to explain its importance to other writers.

“Together, the authors and publishers send a message to the companies of the AI: you are not above the law, and our intellectual property is not yours for taking,” she wrote.

Alsup also declared Thursday in the courtroom that he planned to leave the bench by the end of the year. President Bill Clinton appointed him for the federal bench in 1999.

–

The AP technology writer Matt O’Brien contributed to this history of Providence, Rhode Island.