Were Korean authors' books used to train Claude? - The Korea Herald

Books by South Korean authors, including Nobel Prize winner Han Kang and Hwang Sok-yong, are among the works listed as eligible for compensation in a copyright settlement involving Anthropic, the US company that makes the AI chatbot Claude.

The Literature Translation Institute of Korea said it had found at least 21 English-language editions of books by Korean authors on the list of works eligible for compensation under the settlement.

The institute has informed the authors involved. It said it would work with the relevant copyright organizations to help them receive fair compensation, including by supporting legal action when needed.

In 2024, Anthropic was sued in a group lawsuit called Bartz, et al. v. Anthropic PBC. Three authors accused the company of downloading copyrighted books from pirate websites to train its AI models. A US federal court ruled that Anthropic's use of books to train its AI models was fair use, but that downloading pirated copies of the books was not. The case was settled for $1.5 billion. A federal judge gave final approval in July, making it the largest known settlement in a US copyright case. The settlement works out to about $3,000 per work.

According to the Literature Translation Institute of Korea, the books involved in this case include Kim Ae-ran's "My Brilliant Life" (translated by Chi-Young Kim); four books by Kim Young-ha, including "Diary of a Murderer" (translated by Krys Lee); Kim Soom's "One Left" (translated by Bonnie Oh, Bruce Fulton and Ju-Chan Fulton); three books by Shin Kyung-sook, including "Please Look After Mom" (translated by Chi-Young Kim); three books by Han Kang, including "Human Acts" (translated by Deborah Smith); Hwang Sun-mi's "The Hen Who Dreamed She Could Fly" (translated by Chi-Young Kim); and Hwang Sok-yong's "The Guest" (translated by Chun Kyung-ja and Maya West).

Concerns over suspicious large book orders

Reports first appeared overseas about large book purchases that were suspected of being intended for AI training. Similar cases have recently been reported in South Korea.

Kyobo Book Centre, South Korea's largest bookstore operator, said that since early this year, it had received requests from 10 overseas companies, including companies in the US and UK. Each company wanted to buy around 100,000 book titles.

The bookseller turned down the orders because it could not rule out the possibility that the books would be used to train AI models. A company official said the requests were different from usual large orders, which typically come from libraries with clear reasons for buying the books.

"Many of the companies introduced themselves as wholesale book distributors," a Kyobo official said. When asked why they wanted the books and who would receive them, some companies said the books were not intended for AI training but named data-related companies as the final recipients.

Yes24, another major South Korean bookseller, said a US distributor contacted it in late August to ask about buying used books in large quantities. The two sides met once, but the company did not reply to later messages, and there were no further discussions.

Yes24 said it was watching the situation closely and would carefully check the purpose of similar requests before deciding how to respond.

The Korean Publishers Association has started surveying bookstores, distributors, secondhand bookstores and publishing houses to find out whether they have received large purchase requests that may be linked to AI training.

After the survey, the association plans to bring publishing organizations together to develop a shared response, an official said.

The Korea Publishers Society recently issued a statement calling for fair payment when published works are used to train AI systems. By Tuesday, the statement, released on Thursday, had received support from about 200 organizations and around 570 people, including authors, translators, illustrators and publishing workers.

"Without books, there would be no AI," the association said. "Yet the people who create those books do not even know when, how, or which AI systems are using their work. They cannot ask questions, refuse or demand fair compensation."