2026年5月5日 美国东部时间下午1:07 / 哥伦比亚广播公司新闻
作者：玛丽·坎宁安
玛丽·坎宁安是哥伦比亚广播公司财经观察栏目的记者。她曾任职于《60分钟》、哥伦比亚广播公司新闻网及哥伦比亚广播公司24小时新闻频道，为哥伦比亚广播公司新闻实习项目成员。查看完整简介。
一个出版商团体与畅销小说家斯科特·图罗正在起诉Meta及其创始人马克·扎克伯格，指控这家科技巨头使用受版权保护的材料训练Meta的人工智能技术。
这起集体诉讼在纽约联邦法院提起，由图罗以及圣智学习集团、爱思唯尔、阿歇特出版集团、麦克米伦出版公司和麦格劳-希尔出版公司共同发起。原告方指控Meta未经许可，从互联网上抓取了数百万份受版权保护的作品——其中包括“臭名昭著的盗版网站”上的内容——并利用这些内容训练其AI模型套件Llama。
诉讼称，Meta还删除了作品的版权管理信息，以掩盖其利用被盗材料训练AI的事实。
与其他聊天机器人一样，Llama会根据用户提示生成文本输出。诉状称，这款AI工具会复制小说、期刊文章和教科书等原创作品的内容，在某些情况下甚至会逐字复刻。诉讼还称，Llama在回复中会模仿特定作者的个人风格。
原告方表示，Meta的行为剥夺了作者和出版商本应获得的收入。
诉讼将责任归咎于扎克伯格，称他“绕过正常的许可程序，亲自批准并积极鼓励了侵权行为”。
“由于扎克伯格直接参与Meta的AI开发，包括批准Meta AI通过种子下载盗版作品集来训练Llama，扎克伯格的净资产最近已攀升至2000多亿美元，”诉讼文件中写道。
Meta的一位发言人在发给哥伦比亚广播公司新闻的电子邮件中表示，该公司计划“积极应诉这起诉讼”。
“人工智能正在为个人和企业带来变革性的创新、生产力提升和创造力发展，法院也已合理认定，利用受版权保护的材料训练AI可被视为合理使用，”这位发言人在电子邮件中说道。
文学界此前曾就版权问题与AI公司发生过冲突。据《纽约时报》报道，去年在一起案件中，AI聊天机器人Claude的制造商Anthropic同意向数十万作者支付15亿美元和解金，这是历史上规模最大的版权侵权和解款项。
本周二提起诉讼的原告方表示，他们正在寻求损害赔偿。
艾米·皮奇 编辑
Publishers, author Scott Turow accuse Meta and Mark Zuckerberg of training AI on copyrighted works
May 5, 2026 1:07 PM EDT / CBS News
By Mary Cunningham
Mary Cunningham is a reporter for CBS MoneyWatch. She previously worked at “60 Minutes,” CBSNews.com and CBS News 24/7 as part of the CBS News Associate Program. Read Full Bio.
A group of publishers and bestselling novelist Scott Turow are suing Meta and its founder, Mark Zuckerberg, alleging that the tech giant used copyrighted material to train Meta’s artificial intelligence technology.
The class-action lawsuit, filed in a federal court in New York, was brought by Turow and publishers Cengage, Elsevier, Hachette, Macmillan and McGraw-Hill. The plaintiffs allege that Meta scraped millions of copyrighted works from across the internet —including from “notorious pirate sites”— and used the content to train Llama, Meta’s suite of AI models, without permission.
Meta also removed copyright management information from the works to hide the fact that it was training its AI on stolen materials, the lawsuit alleges.
Like other chatbots, Llama generates text outputs in response to user prompts. The complaint claims that the AI tool is reproducing versions of original works from novels, journal articles and textbooks, and in some cases recreating verbatim copies. Llama also mirrors certain authors’ personal style in its responses, according to the lawsuit.
The plaintiffs say Meta’s actions are robbing authors and publishers of revenue they would otherwise receive.
The suit assigns blame to Zuckerberg, claiming he “personally authorized and actively encouraged the infringement” by sidestepping normal licensing procedures.
“As a result of Zuckerberg’s day-to-day involvement in Meta’s AI development, including his authorization for Meta AI to torrent pirate collections to train Llama, Zuckerberg’s net worth recently climbed to over $200 billion,” the lawsuit says.
A Meta spokesperson told CBS News in an email that the company plans to “fight this lawsuit aggressively.”
“AI is powering transformative innovations, productivity and creativity for individuals and companies, and courts have rightly found that training AI on copyrighted material can qualify as fair use,” the spokesperson said in an email.
The literary world has previously clashed with AI companies over copyright issues. In a case last year, Anthropic, maker of the AI chatbot Claude, agreed to settle with hundreds of thousands of authors for $1.5 billion, the largest payout for copyright infringement in history, according to The New York Times.
The plaintiffs in Tuesday’s suit said they are seeking damages.
Edited by Aimee Picchi
发表回复