仅事实
Anthropic与美国作家达成AI版权侵权诉讼和解
摘要
Anthropic已与美国作家就其使用盗版书籍训练AI模型的指控达成和解，避免了一场可能代价高昂的审判。
人工智能公司Anthropic已与美国作家提起的集体诉讼达成和解。原告指控该公司使用盗版书籍训练其AI助手Claude。和解条款未予披露，但该协议使Anthropic避免了一场可能导致巨额经济赔偿的审判。
该诉讼由作家Andrea Bartz、Charles Graeber和Kirk Wallace Johnson发起，指控Anthropic从网络“影子图书馆”下载了数百万版权作品以开发其AI模型。6月，美国地方法官William Alsup裁定，使用合法获得的书籍训练AI模型属于合理使用，但获取盗版副本则不属于。他指出，Anthropic可能下载了多达七百万本盗版书籍，可能使公司面临重大赔偿责任。
和解预计将于9月3日最终确定。作家代理律师Justin Nelson称该协议为“历史性”，对所有集体成员都有利。Anthropic拒绝对和解事项发表评论。
此案是针对OpenAI、微软和Meta等AI公司使用版权材料训练AI模型的多起类似诉讼之一。此次和解结果可能影响人工智能与版权法这一快速发展的领域未来的诉讼走向。
来源
The Verge事实核查
使用外部来源和数据库核实文章的事实。
✓
Confirmed
Anthropic has settled a class action lawsuit filed by U.S. authors who alleged that the company used pirated books to train its AI assistant, Claude.
✓
Confirmed
The terms of the settlement remain undisclosed.
✓
Confirmed
The lawsuit accused Anthropic of downloading millions of copyrighted works from online 'shadow libraries' to develop its AI models.
✓
Confirmed
U.S. District Judge William Alsup ruled that while training AI models on legally obtained books constitutes fair use, the acquisition of pirated copies does not.
!
Partly Confirmed
Anthropic may have downloaded up to seven million pirated books, potentially exposing the company to significant damages.
✓
Confirmed
The settlement is expected to be finalized on September 3.
✓
Confirmed
Justin Nelson, an attorney for the authors, described the agreement as 'historic' and beneficial for all class members.
✓
Confirmed