[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-anthropic-sony-warner-music-copyright-lawsuit":3,"topics-all":38,"news-related-264b5334-0465-48b8-ad81-b2b7b39d1a3f":57},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"264b5334-0465-48b8-ad81-b2b7b39d1a3f","Anthropic 被索尼华纳告上法庭：两万首歌喂出来的 Claude 还要赔多少","索尼音乐和华纳音乐在纽约联邦法院起诉 Anthropic,指控其用 BitTorrent 下载盗版图书和歌词训练 Claude,索赔最高每部 15 万美元。这场诉讼复制了 Bartz 案的打法(7 月已判赔 15 亿美元),只打数据获取方式,不打训练是否合理使用。","8 月底的一个周五深夜，48 页起诉书递进纽约联邦法院。索尼音乐和华纳音乐把 Anthropic 连同创始人 Dario Amodei 和 Benjamin Mann 一起告上法庭，索赔金额明明白白写在封面上：每部侵权作品最高 15 万美元，每次故意删除版权信息再加 2.5 万美元。如果法院按上限判，数字会是数十亿美元。这不是 Anthropic 第一次因为训练数据被起诉，但这一次，律师手里握着 Bartz v Anthropic 那张已经判下来的牌。\n\n## 告的是什么\n\n根据 Axios 拿到的诉状，唱片公司指控 Anthropic\"明目张胆地、系统性地\"做了三件事来给 Claude 喂数据：第一，创始人之一 Benjamin Mann 用 BitTorrent 下载了超过 500 万本盗版图书；第二，Anthropic 的员工从 Pirate Library Mirror（俗称\"海盗图书馆镜像\"）下载了至少 200 万本盗版书；第三，公司从 MusixMatch、LyricFind 这些本来已经付费授权给唱片公司的歌词站上抓取歌词。这次的标的不是\"几千首\"而是\"数万首\"——相比今年早些时候 BMG 起诉 Anthropic 时只主张 493 部作品被侵权，规模直接放大了一个数量级。\n\n## 索赔怎么算\n\n诉状的赔偿条款是这么写的：每一部被侵权的版权作品最高 15 万美元，再加上每次\"可识别的版权信息被删除\"的情形最高 2.5 万美元。按 Sony\u002FWarner 现在列的\"数万首\"作品计算，光这一项就已经是天文数字，更别提还要叠加故意删除版权元数据的惩罚性赔偿。唱片公司在诉状里直接用了\"历史上规模最大、最明目张胆的知识产权盗窃之一\"这种话，几乎是把 Bartz 案判决书里的话换了个主语。\n\n## 为什么这次不一样\n\n关键不在\"AI 公司用盗版数据训练\"这个罪名——Bartz v Anthropic 在 7 月 20 日的判决已经明确：训练本身是 fair use，这条法律阵地 Anthropic 守住了。但同一份判决也认定：用 BitTorrent 获取盗版数据是版权侵权，按这条 Anthropic 已经赔了 15 亿美元。\"怎么拿到训练数据\"和\"训练数据怎么用\"是两条战线，Anthropic 已经输了一条。\n\nSony\u002FWarner 这场诉讼，几乎就是把 Bartz 案那套打法原样复制到音乐领域：同样主张 Anthropic 用 BitTorrent 获取数据，同样要求每部作品 15 万美元上限赔偿。唱片公司的算盘很清晰——既然法庭已经认定\"BitTorrent 下载是侵权\"的判例，那就顺着已经跑通的路走。不确定性大幅下降。\n\n## 行业后果\n\n这意味着接下来一年里，所有靠爬虫、BT、镜像站喂过数据的 AI 公司都要重新算账。法律上\"训练\"和\"获取\"被切开——可以继续合法地用公开数据集训练，但必须在数据来源环节做尽职调查。如果抓取路径里出现过 BitTorrent 客户端、Pirate Library Mirror 这类明显有问题的源头，哪怕模型本身没问题，公司还是为\"脏数据\"买单。\n\n更深的行业影响是版权圈的预期：BMG 看到这条路能走通，后面环球音乐、EMI 这些大体量玩家排队进场的概率很高。AI 行业引以为傲的\"开放数据集\"叙事，在 2026 年的法庭上已经快撑不住了。\n\n所以回到开头那个问题——Claude 还要赔多少？数字不重要，重要的是法律框架已经定型。AI 行业必须在 2027 年之前把数据获取的合规链彻底重建，否则下一波诉讼会沿着 Bartz + Sony\u002FWarner 这条已经铺好的路直接撞过来。","https:\u002F\u002Fwww.solidot.org\u002Fstory?sid=85234","d59894d3-308e-4fd8-8865-86dc1eeac4a2",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"c33b1bbc-d6ce-4f61-9d5d-1a0704a6a09b","ai-policy",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",{"id":19,"name":20,"slug":20,"description":14,"color":14},"23544f6a-eea1-4f05-aa8d-749ca862d5d2","anthropic",{"id":22,"name":23,"slug":23,"description":14,"color":14},"01598627-1ea6-4b27-a5d8-874971571a71","llm",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"f6359006-7ad1-4158-8f09-4cd6700c1c10","en","Sony and Warner Sue Anthropic Over Claude Training Data","Sony and Warner sued Anthropic in New York over BitTorrent book downloads and lyric scraping for Claude, seeking up to 50K per work. The suit copies the Bartz playbook (settled July 20 for .5B) and targets acquisition, not training.","A 48-page copyright complaint landed in a New York federal court late on a Friday in late August. Sony Music and Warner Music Group named Anthropic — together with co-founders Dario Amodei and Benjamin Mann — as defendants, and put a number on page one: up to $150,000 per infringed work, plus up to $25,000 each time identifying copyright information was stripped. At statutory maximums, the math runs into billions. Anthropic has been sued over training data before, but this time the plaintiffs walked in with a precedent already on the books: Bartz v Anthropic, settled for $1.5 billion on July 20 after a court ruled that BitTorrent-based acquisition of training material was copyright infringement, even while training itself remained fair use.\n\n## What the Complaint Alleges\n\nAccording to details first reported by Axios, the publishers accuse Anthropic of running a three-track pipeline to feed Claude. Co-founder Benjamin Mann allegedly used BitTorrent to download over five million pirated books. Anthropic employees pulled at least two million more from Pirate Library Mirror, an open repository of pirated texts. The company also allegedly scraped lyrics from MusixMatch and LyricFind — services that had paid licensing fees to the very publishers now suing. The headline number is scale: tens of thousands of musical compositions. BMG's earlier Anthropic suit covered 493 works. Sony and Warner have stepped that up by orders of magnitude.\n\n## How the Damages Math Works\n\nEach infringed work carries a statutory ceiling of $150,000; each metadata-stripping incident adds another $25,000. Multiply by tens of thousands of compositions and the number blows past anything a quiet settlement can absorb. The complaint's framing — \"one of the largest and most blatant ongoing thefts of intellectual property in history\" — reads almost like an edit of language the Bartz court used about the same defendant in a different medium.\n\n## Why This Case Is Different\n\nThe legal leverage here is not new. Bartz v Anthropic already established that training an AI on copyrighted works is fair use — Anthropic won that point. What Anthropic lost in Bartz was the acquisition channel: courts held that obtaining those works via BitTorrent was itself infringement, and the company settled for $1.5 billion. The Sony\u002FWarner complaint is the same playbook applied to music — same defendants, same BitTorrent acquisition theory, same $150,000-per-work demand. Plaintiffs are not asking courts to reinvent the law; they are asking courts to extend an existing ruling into a new content domain.\n\n## Industry Implications\n\nThe short-term fallout is sharper liability lines for any AI lab whose data trail includes torrents, pirate mirrors, or unlicensed scraping. The worlds of training and acquisition have been cleanly separated, and the second one is now the active battlefield. Companies can keep training on legitimately sourced public data, but the diligence bar on provenance just moved.\n\nIf Sony and Warner can ride Bartz to a meaningful win — even a settlement — Universal Music, EMI, and the rest of the major-label bench will line up to file similar suits. The \"open dataset\" narrative that has powered AI lab marketing since 2023 is running out of courtroom oxygen.\n\nThe headline number matters less than the legal framework now in place. The AI industry has until roughly 2027 to rebuild its data-acquisition compliance chain, or the next wave of lawsuits will land on the exact road Bartz and Sony\u002FWarner have already paved.","anthropic-sony-warner-music-copyright-lawsuit","2026-09-05T00:00:00Z","2026-09-05T09:06:31.932190Z","2026-09-05T09:06:31.932201Z",true,"agent",107,[39,48],{"slug":40,"tag_slug":40,"title_zh":41,"title_en":42,"intro_zh":43,"intro_en":44,"id":45,"is_active":35,"created_at":46,"modified_at":47},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":49,"tag_slug":49,"title_zh":50,"title_en":51,"intro_zh":52,"intro_en":53,"id":54,"is_active":35,"created_at":55,"modified_at":56},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":58},[59,64,69,74,79,84],{"id":60,"title":61,"news_slug":62,"published_at":63},"d3e055f9-1fbe-4d91-9eca-f336b930be80","索尼华纳起诉 Anthropic:每首歌索赔 15 万美元,可能拖出又一份 10 亿美元和解","sony-warner-anthropic-billion-dollar-lawsuit","2026-08-31T11:00:00+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"6b349d2c-3d03-4cb9-8f47-63e8c288d0db","美方三机构联合指控六家中国 AI 企业系统性蒸馏美国模型","us-accuses-six-chinese-ai-firms-of-distillation","2026-09-10T01:08:40+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"d17a841b-abca-46e0-80e4-d955f1c837ba","亚马逊 VGT3 仓库曝光:一天拆掉上千本书,只为给 AI 模型喂语料","amazon-vgt3-warehouse-ai-training-books","2026-09-07T03:30:00+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"1942b07b-f794-42b1-b944-ca6b32d4ae16","四大 AI 模型同日集体掉线:OpenAI\u002FClaude 官方确认,Gemini\u002FGrok 表面沉默","four-ai-models-overlapping-outage-sept-2026","2026-09-06T08:00:00+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"2fc4f696-a697-498c-b9d5-28250bfeaa79","ChatGPT 进欧盟 VLOP 名单:OpenAI 第一次要为生成式 AI 内容负全责","chatgpt-eu-vlop-dsa-first-ai-platform-rules","2026-09-05T07:00:00+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"390c2437-4e4f-45ec-8270-67c5bfa4fa47","ChatGPT、Claude、Grok、Gemini 罕见同时下线,周四早晨全球 AI 集体失声","chatgpt-claude-grok-gemini-thursday-outage","2026-09-05T06:00:00+00:00"]