Nihon Hanyu Supplying Books to Anthropic Without Permission Sparks Controversy; Publishers Call It Unacceptable

Overview

A subsidiary of Nihon Hansha Holdings, one of Japan's major publishing distribution groups, has been exposed for selling Japanese books to the US artificial intelligence (AI) developer Anthropic without permission, sparking strong discontent within the publishing industry. According to industry insiders, some books were mass-scanned and used to train AI models, an act widely viewed as a serious infringement of copyright law. Several publishing houses have stated they "cannot accept" such practices.

Background of the Incident and Key Points of Controversy

Nihon Hansha, one of Japan's largest publishing distribution companies, has a subsidiary involved in supplying Anthropic with large volumes of book resources. These books were used directly for AI dataset training without authorization from publishers, resulting in the uncompensated use of intellectual property by original authors and publishing entities. This has triggered deep concerns within the publishing industry regarding the boundaries of "fair use" and has been widely criticized as a disregard for intellectual property rights.

"When books are sold in bookstores, there is no requirement for buyers to confirm whether the publisher agrees," noted one publishing industry figure, "but using books for AI training is fundamentally different from ordinary sales."

Industry Reactions and Legal Challenges

The Japan Books Publishers Association (Nihon Shuppan Hanbai Kyogyo Kyokai) has raised questions about this incident but has not yet received a clear response. The industry generally believes that if such behavior is allowed to continue unchecked, it will severely dampen creative enthusiasm and undermine the economic foundation of the publishing sector. Legal experts analyze that while existing copyright law lacks clear definitions regarding AI data scraping, "unauthorized copying plus commercial use" clearly constitutes infringement.

"Allowing AI companies to access all books for free is equivalent to making society pay for their benefits," stated one lawyer.

Long-Term Impact and Future Directions

This incident relates not only to the conduct of a single enterprise but could also reshape the rules of interaction between AI and the publishing industry. If industry standards are not established in a timely manner, similar incidents may occur frequently, posing systemic risks to the publishing ecosystem. Currently, the association is coordinating among its member units to prepare collective legal action and policy advocacy to protect their rights and promote improvements in legislation.