An AI bot used their words. Now, authors are demanding compensation
In Andrea Bartz’s book “The Lost Night,” the main character uncovers dark truths about an unsettling event and delves into themes of memory and technology. So, it was an ironic twist of fate when Bartz discovered that her debut thriller had been pirated to teach an AI chatbot named Claude.
The battle between artificial intelligence and artists’ rights raged on in a San Francisco courtroom, where Bartz and a class of authors brought their copyright suit against Anthropic, the Amazon-backed company behind Claude. Anthropic agreed to pay a record-setting settlement of $1.5 billion. But with the federal judge overseeing the case lambasting the settlement, what seemed like an end to this legal battle on the frontier of technology and art may be just the beginning.
Authors take AI to court
Bartz and two other writers — Charles Graeber, who wrote “The Good Nurse,” and Kirk Wallace Johnson, author of “The Feather Thief” — sued Anthropic in 2024 on behalf of themselves and the class, accusing the company of stealing their books without permission or payment to train Claude’s large language model, or LLM.
Anthropic balked, citing the legal doctrine known as fair use, which allows limited use of copyrighted material without permission, for such purposes as teaching, scholarship and research.
Anthropic said it was not copying the books, but instead using them to train its chatbot.
In U.S. copyright law, using another person’s art without their permission is allowed if the new work takes on a different form rather than copies the work.
In a landmark decision in June, U.S. District Judge William Alsup sided with Anthropic, comparing the way it trains AI to how a human reads books and then writes about new ideas inspired by the books.
“Like any reader aspiring to be a writer, Anthropic’s LLMs trained upon works not to race ahead and replicate or supplant them — but to turn a hard corner and create something different,” the judge wrote. “The technology at issue was among the most transformative many of us will see in our lifetimes.”
However, there was a caveat: If a company uses pirated downloads or unlicensed copies, the judge said, it may not be protected by fair use.
Judge slams largest settlement in U.S. copyright history
Because Anthropic acquired millions of pirated books through online libraries such as Pirate Library Mirror and Library Genesis, the judge set the trial regarding the legality of the books’ provenances for Dec. 1.
But then Anthropic offered the largest settlement in U.S. copyright history to the class of authors for about 500,000 books.
“The June ruling, which found that the fundamental legal principle that AI training on copyrighted works is transformative fair use, remains unchanged,” Sally Aldous, an Anthropic spokesperson, told Straight Arrow News. “This settlement simply resolves disputes about how certain materials were obtained.”
Whether the case goes to trial is undecided. The judge said earlier this month that the settlement was “nowhere close” to being approved and is full of holes that may not protect authors in the future. He scheduled the next hearing to review the settlement for Sept. 25.
AI suits prompt new questions
In an era when the future of AI and creativity is at stake, 40 similar copyright suits are stacking up in U.S. courts. Tech companies have held fast to their stance that they are not paying for creative material because the information is public and they are not copying the works.
But many authors, artists, journalists and musicians say otherwise.
“All people who make content that can be digitized are being taken advantage of,” David Friend, an author and contributing editor at Vanity Fair, told SAN.
Experts believe the class-action suit against Anthropic could set the tone for other copyright cases, in which tech companies will pay fees to the holders of copyrighted works.
For now, however, it leaves questions unanswered: How is creativity defined in this new era? How should it be protected? What ethics are involved? Should AI be treated like human beings?
“This is people’s licenses and royalties,” said F. Elusio Zafar, owner and founder of Creation Rights, a company that protects artists’ intellectual property. “This is what they work their whole life to make an impact on society — and that’s what contributes to our cultural knowledge. That’s being threatened now by AI companies.”
How tech companies acquire books also matters, Zafar told SAN.
“Even when AI use of content is considered fair, the way that content is acquired can create enormous legal risk,” he said. “In Anthropic’s case, the AI training itself was legal, but the books were sourced from pirate ‘shadow libraries,’ leaving the company exposed to potential damages in the billions. This highlights that provenance and legal clearance are just as important as the AI’s output. Meta may face the same challenge as they also used ‘shadow libraries’ for training.”
First-of-its-kind settlement agreement
Attempting to avoid trial, Anthropic offered the authors the largest settlement in U.S. copyright case history: $1.5 billion. The amount comes out to $3,000 each for about 500,000 authors.
“Prior to this settlement agreement, tech companies were saying they weren’t violating the law, they said it was fair use,” Zafar told SAN. “This is the first time where you’re seeing that it’s actually taking people’s work without compensating them. It sets a standard of accountability and puts a price tag on something that’s been intangible. … Financially no one could give an actual value of risk analysis or loss analysis. This is the first time there is a settlement number. They’re paying out $1.5 billion for a specific category in the creative industry, which is authors and books.”
Both the plaintiffs’ attorneys and Anthropic seemed pleased with the settlement.
“In June, the District Court issued a landmark ruling on AI development and copyrights law, finding that Anthropic’s approach to training AI models constitutes fair use,” Aparna Sridhar, Anthropic’s deputy general counsel, said in a statement. The settlement “will resolve the plaintiffs’ remaining legacy claims. We remain committed to developing safe AI systems that help people and organizations extend their capabilities, advance scientific discovery, and solve complex problems.”
“This landmark settlement far surpasses any other known copyright recovery,” Justin Nelson of the law firm Susman Godfrey, co-lead counsel for the plaintiffs, said in a statement. “It is the first of its kind in the AI era, and sets a precedent requiring AI companies to pay copyright owners.”
The judge, on the other hand, took issue with how the authors’ claims process would work to ensure everyone who is eligible receives a payout. He said groups such as the Authors Guild and the Association of American Publishers could mislead authors about what signing the settlement entails.
In addition, the settlement would end the possibility of trial, where an important legal precedent could be established around pirating works of art.
“I think we are moving into a new age of intellectual property protection,” Zafar said. “AI companies are fighting artists about their rights. If they just include their rights, then both parties will get more ownership, more creativity, more results.”
The post An AI bot used their words. Now, authors are demanding compensation appeared first on Straight Arrow News.
Ella Rae Greene, Editor In Chief
