ReadingResearchRadarInvestment framework
Sign in / Sign up
← Back to reading
TopicDisconfirmAI EditorialUpdated 2026-10-26 11:45

Anthropic版权案

4 steps2 pieces4 Sources

美国联邦法院在Bartz v. Anthropic案中确立了AI版权责任的“双轨制”：模型训练本身可构成合理使用，但从盗版影子图书馆获取数据仍属侵权。这一裁决终结了“先训练后和解”的策略，迫使Anthropic支付15亿美元和解金。然而，随着2026年10月猫王音乐版权方起诉Anthropic未经授权使用274首歌曲训练Claude并索赔5000万美元，争议焦点已从书籍转向音乐领域。核心分歧在于：当AI输出与原作品高度相似时，“转换性使用”的抗辩是否依然有效？读者需关注法院对音乐类训练数据的最新认定及和解款项的实际分配进度。

How it got here

  1. 2025-06-23

    美国加州北区联邦地区法院法官William Alsup作出即决判决，认定Anthropic将合法购买的印刷书籍转换为数字副本用于训练Claude属于合理使用，但指出从盗版来源获取数据不构成合理使用。

    ailawwiki.com ↗
  2. 2026-07-20

    法官Araceli Martínez-Olguín最终批准Anthropic与作者群体的15亿美元集体和解协议，案件被驳回，法院保留对和解管理的管辖权。

    ailawwiki.com ↗
  3. 2026-09-04

    美国作家协会报告称，和解通知中存在出版商与作者之间的索赔分配错误，并启动了直接谈判或提交特别主事官（Special Master）的程序以解决争议。

    ailawwiki.com ↗
  4. 2026-10-23

    猫王音乐版权持有者（包括Authentic Brands Group）在美国联邦法院起诉Anthropic，指控其未经许可使用274首猫王歌曲训练Claude，要求赔偿约5000万美元。

    hollywoodreporter.com ↗

What is still unsettled

  • 法院是否会沿用Bartz案的逻辑，认定音乐歌词的训练也属于“转换性使用”，还是因Claude能生成高度相似的歌词而判定侵权？
  • 15亿美元和解款中，针对存在分配争议的书籍，最终支付给作者和出版商的具体比例何时确定？
  • Anthropic在猫王案中提出的“护栏”措施是否足以作为免责理由，还是被视为明知故犯的证据？
Sign in / Sign up
ReadingResearchRadarInvestment framework

Sources

  • ailawwiki.comAI Law Wiki: Bartz v. Anthropic PBC Case Summary↗
  • hollywoodreporter.comThe Hollywood Reporter: Anthropic Sued for Using Elvis Presley Music↗
  • latimes.comLos Angeles Times: Elvis rights holder sues Anthropic↗
  • techorange.comTechOrange: Analysis of Bartz v. Anthropic Ruling↗

Coverage in this topic

  • Elvis Music Copyright Holders Sue Anthropic: Claim Unauthorized Training of Claude on 274 Songs, Seek $50 Million2026-10-26News

    Copyright holders of Elvis Presley's music have sued Anthropic in US federal court, alleging the company used 274 Elvis songs to train its AI model Claude without permission and claiming the AI generates similar lyrics, seeking $50 million in damages.

  • Former NCC Chairman Chan Ting-yi Analyzes Anthropic Copyright Case: Court Splits AI Training into Two Independent Acts, Ending the 'Train First, Settle Later' Strategy2026-10-21News

    Chan Ting-yi uses the ruling in Bartz v. Anthropic as a starting point to analyze how Federal Judge William Alsup deconstructed the AI model training process into two independent legal acts: 'frontend data acquisition' and 'backend model training,' applying different standards to each. This framework of behavioral decomposition implies that even if the training itself constitutes fair use, obtaining data from pirate shadow libraries remains illegal. The author argues this fundamentally shifts the bargaining power between AI companies and publishing media, predicting that collective licensing organizations, synthetic data, and proactive defense technologies will become the trends of the next stage.