法官裁定Anthropic使用非法取得的书籍训练Claude
在Bartz v. Anthropic案中,法官William Alsup裁定,Anthropic非法抓取并盗版使用十多位斯坦福教授的作品,为Claude建立训练资料库。法官同时认定,使用合法取得的书籍训练属于合理使用。15亿美元和解向近50万部登记作品的作者和出版商提供补偿,但Tobias Wolff等28位作者认为补偿不足。Wolff和Chang-rae Lee选择退出该和解,并在Cruz v. Anthropic案中另行提出诉求。
该裁决据报确立了针对盗版书籍训练资料库的现行规则,同时将使用合法取得的书籍训练视为合理使用。原报道指出,大规模抓取数据用于AI模型训练的法律前景仍不确定,作者团体正在推动更多诉讼及授权相关行动。
AI giant Anthropic illegally scraped and pirated the work of over 10 Stanford professors to train its Claude models, a federal judge ruled in a May 2026 lawsuit.
Anthropic previously trained Claude on illegally obtained texts from professors Tobias Wolff, Abraham Verghese, Chang-Rae Lee, Daniel Mason, Hayagreeva Rao and Myra Strober. In Bartz v. Anthropic, which the Authors Alliance calls the largest copyright settlement involving an AI company, Judge William Alsup ruled that Anthropic could no longer pirate books to create a permanent central library.
While blocking the piracy of work to train AI, Alsup justified the legal acquisition of books to train Claude’s Large Language Models (LLMs) as fair use. He compared AI training to how humans can read, internalize and remember a book’s themes and style. “Like any reader aspiring to be a writer, Anthropic’s LLMs trained upon works not to race ahead and replicate or supplant them — but to turn a hard corner and create something different,” he wrote.
The $1.5 billion settlement went to authors and publishers of nearly 500,000 registered works. Wolff and 27 other authors say the settlement isn’t enough.
#Anthropic#Claude#Bartz v. Anthropic#Cruz v. Anthropic#Stanford University