Japanese publishers protest Anthropic's large-scale book purchases for AI training

Japanese publishers have expressed alarm over Anthropic's bulk purchasing of books for use in training its AI models, marking the latest friction point between the technology sector and content creators over data sourcing practices.

The California-based AI firm's practice of acquiring large numbers of Japanese publications has drawn criticism from industry groups who argue the approach circumvents proper licensing arrangements and fails to compensate authors and publishers adequately. The backlash reflects broader concerns across Asia's creative industries about how generative AI systems are being developed and trained without explicit consent from rights holders.

Publishers in Japan have questioned whether mass book purchases constitute fair use of copyrighted material, particularly when the works are subsequently used to generate commercial AI products. The dispute mirrors similar confrontations between content creators and AI companies in Western markets, where authors and publishers have challenged the legality of using published works without permission or compensation.

AnthropIc has not publicly commented on the Japanese publishers' complaints. The company has previously stated that its data sourcing practices comply with applicable laws, though it has faced multiple lawsuits in the United States from authors alleging copyright infringement related to its AI training methods.

The controversy underscores the regulatory challenges facing AI developers in Asia, where countries are still formulating policies to balance innovation with intellectual property protections and creator rights.