Han Kang's Works Also Targeted... Anthropic Used 21 Books for AI Training Without Permission

It has been revealed that the U.S. artificial intelligence (AI) company Anthropic used English-translated books by South Korean authors, including Han Kang and Hwang Sok-yong, without authorization to train its AI chatbot, Claude.

According to the Literature Translation Institute of Korea (LTI Korea), the institute discovered that 21 books by South Korean authors were included on a compensation target book list posted on a copyright settlement website related to an Anthropic lawsuit, and notified the affected authors.

Anthropic had previously faced a class-action lawsuit over allegations of illegally downloading and using copyrighted materials to train its AI models.

A U.S. court acknowledged about 500,000 instances of copyright infringement, and the lawsuit concluded in July of this year with Anthropic agreeing to pay a total of 1.5 billion dollars, approximately 2 trillion won in South Korean currency.

The compensation per work is about 3,000 dollars, or around 4.35 million won, and domestic authors can also receive compensation if they participate in filing claims for it.

According to LTI Korea, the works whose contents were downloaded without authorization by Anthropic include Kim Ae-ran's "My Palpitating Life," Kim Young-ha's "Empire of Glow" and three other works, Kim Sum's "One Person," Shin Kyung-sook's "Please Look After Mom" and two other works, Han Kang's "Human Acts" and two other works, Hwang Sun-mi's "The Hen Who Dreamed She Could Fly," and Hwang Sok-yong's "The Guest."

Regarding moves by overseas AI companies to purchase large quantities of books suspected of being used for training, the Korean Publishers Association also released an appeal yesterday (October 16) calling for fair compensation.

(Photo: AP, Yonhap News)
※ Please note: This article was translated by AI and may contain errors.