EVENT DOSSIER
According to TechCrunch, the legal issues surrounding the use of copyrighted book data to train artificial intelligence models are extremely complex. No clear consensus has yet been reached; the main争议 points lie in whether copyright laws permit such large-scale data processing and whether there are legitimate exemption mechanisms for such use. This matter is still in the stage of legal discussion, and specific compliance requirements need to be further clarified based on individual cases and judicial precedents.
TechCrunch 记者 Amanda Silberling 于 2026 年 8 月 23 日报道，训练 AI 模型使用受版权保护书籍的合法性问题复杂。尽管大多数作者未知情便贡献了数据，但去年法官 William Alsup 裁定 Anthropic 使用这些书籍训练 AI 合法，仅因其从非法影子图书馆盗版书籍而处以 15 亿美元罚款。律师 Cathy Gellis 指出，该判决对 AI 公司有利，因为版权法核心在于“复制”而非“使用”，且自 1976 年未更新，法官需依据过时指南解释新问题。资深律师 Jason Henderson 表示，法律尚未跟上 AI 模型训练海量数据的现实，争议常围绕“合理使用”是否构成足够“转化性”展开。