
2026年3月,十三家主要出版商起诉了一个旨在汇集全世界所有图书的网站。此案折射出当下图书盗版的现状,也说明了为何要遏制这一现象如此困难。下文简要梳理事件经过、问题严重程度,以及为何下架、网站屏蔽和诉讼正日益由自动化来处理。
2026年3月6日,美国出版商协会(AAP)对Anna’s Archive提起诉讼,共有十三家成员参与,包括Penguin Random House、HarperCollins、Hachette、Simon & Schuster、Elsevier和Wiley。该案名为《Apress Media 等诉 Anna’s Archive》,已提交至曼哈顿联邦法院。
原告主张对方构成蓄意侵犯版权。他们指出,该平台以商业化方式经营图书盗版,并利用“捐赠”充当会员费。起诉书中列举了130部作品,涵盖从Jonathan Franzen的“The Corrections”到“PowerPoint 2019 For Dummies”。AAP负责人Maria Pallante称其为“肆无忌惮的盗版行径”。由于涉案个人身份匿名,诉状仅称他们为“Does 1-10”。
那么,当前图书盗版的规模如何?起诉书给出了两个令人瞩目的数字:
- 超过6300万本图书和9500万篇论文,约1 PB数据。
- 以该平台截至2026年3月3日的30日均值计算,每天大约有763,000次下载。
有必要为这些数字提供一些背景:它们基于原告和盗版平台自身的统计,属于主张而非经核实的总量。此外,一次下载并不必然等同于一次销售流失,因为有些人原本就不会为该书付费。
但损害确实存在。这类资源库常被称为影子图书馆,但提起诉讼的一方不接受这一说法。正规图书馆外借的是其合法购买的馆藏;与此同时,每一份免费副本都可能意味着销量流失,以及作者版税的损失。尽管数字版权管理(DRM)锁可阻碍简单的复制行为,但执意上传的人仍可将其移除,这也正是执法通常在内容已泄露之后才启动的原因。
行业如何应对:法院、注册机构与搜索
行业主要使用三种工具来打击图书盗版,而每一种都有缺陷。
该案进展迅速。2026年5月19日,因运营者未出庭,杰德·拉科夫法官作出缺席判决。他针对130部作品,按每部作品15万美元判赔法定最高损害赔偿,总额达1950万美元。
更具意义的是法院禁令,因为它要求包括Cloudflare在内的二十余家注册机构、注册商和托管商停用剩余域名。这是在源头实施域名屏蔽。然而,8月时TorrentFreak报道称,该盗版平台仍可通过由美国境外注册商管理的.gl、.pk和.gd地址访问。对Cloudflare等服务商施加的法律压力,正迫使运营者不断更换其基础设施。
搜索也是打击图书盗版的一条战线。截至2025年底,Google已验证与其平台地址相关的7.49亿次移除请求,约占自2012年以来因版权理由被举报URL总数的5%。尽管如此,当以名称进行搜索时,主域名仍然排在前列。简言之,将网站从搜索引擎下架,能移除单本图书页面,却并不能隐藏主要入口。
图书盗版新闻:执法走向自动化,AI参与其中
关键变化在于,如今的执法在很大程度上已实现自动化。到2025年底,版权方每周仅针对这一家平台就报告约1000万条新URL。人工团队已无法应对如此庞大的数量。荷兰公司Linkbusters代表出版商提交数百万条自动化下架请求,因此如今许多DMCA通知都是自动生成的。
法院命令沿袭了唱片公司此前起诉该网站时的模式,该案同样点名了其中一些相同的中介机构。核查屏蔽是否生效需要切实投入,版权方和反盗版团队往往会购买代理访问权限,像本地访客那样查看某个区域的盗版站点版本。
AI的介入带来了新的压力。起诉书称,运营者向中国、俄罗斯及其他国家的公司提供对超过1.4亿篇文本的高速访问,其中一些公司正在开发语言模型。该网站还鼓励开发者进行“企业级捐赠”,报道称支付20万美元加密货币即可获得完整访问权限。
起诉书援引加州一家法院的认定称,Meta在开发其Llama模型时下载了该收藏。原告主张,这降低了经授权的AI训练数据的市场需求。
请注意,这些均为指控;由于运营者在被法院传唤时未答辩,因此法院作出缺席判决,相关主张并未在庭审中经过检验。
图书盗版最直接的受害者是作者。免费副本的存在会减少销量,而且盗版内容还可能在未经许可的情况下进入AI系统。如果你写作、出版或阅读图书,以下几点或许有帮助:
- 对于作者:与你的合作出版社确认是否已有监测服务在代你的书籍提交下架请求。Google的报告指出,已有超过1,000位作者和出版商就该平台发出了通知。
- 对于读者:不要盗版图书;请通过合法书店购买,或使用图书馆应用借阅。
- 对于所有人:关注AI授权许可的争论。这可能是围绕图书的下一场大博弈。
打击图书盗版的斗争远未结束。尽管法院已作出巨额判决,但美国的命令可能无法在美国境外的注册机构强制执行。这一漏洞很可能引发对更严格域名屏蔽规则的呼声,尤其是在站点屏蔽方面。进展将是循序渐进的,而每一步都会让盗版资源更难找到。
Markus居住于加利福尼亚州旧金山,是Good e-Reader的电子游戏与音频专家!他对新型电子阅读器和平板电脑以及游戏兴趣浓厚。
本文来源:https://goodereader.com/blog/electronic-readers/book-piracy-is-going-automated-as-publishers-fight-back
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。
In March 2026, thirteen major publishers sued a website that aims to gather every book in the world. The case illustrates the current state of book piracy and explains why it is so difficult to stop. The following offers a simple look at what happened, how serious the problem is, and why takedowns, site blocks, and lawsuits are increasingly handled by automation.
On March 6, 2026, the Association of American Publishers (AAP) filed a lawsuit against Anna’s Archive, with thirteen members joining the suit, including Penguin Random House, HarperCollins, Hachette, Simon & Schuster, Elsevier, and Wiley. The case, named Apress Media et al. v. Anna’s Archive, was filed in the federal court in Manhattan.
The plaintiffs claim willful copyright infringement. They argue that the operation runs book piracy as a commercial enterprise, using “donations” which function as memberships. The complaint includes 130 examples of works, ranging from Jonathan Franzen’s The Corrections to PowerPoint 2019 For Dummies. Maria Pallante, chief of AAP, described it as a “brazen pirate operation.” Since the individuals involved are anonymous, the lawsuit refers to them only as “Does 1-10”.
So what is the current scale of book piracy? The complaint gives two remarkable figures :
More than 63 million books and 95 million papers, about one petabyte of data.
Roughly 763,000 downloads a day, based on the operation’s own 30-day average on March 3, 2026.
It’s important to have some context for these figures, since they are based on the plaintiffs’ figures and the pirate platform’s own counts, meaning they are claims, not verified totals. Furthermore, a download does not always represent a lost sale because some people would never have paid for the book.
The damage is still there. Such collections are often called shadow libraries, but the people who brought the case do not accept that term. Genuine libraries lend out books that they have purchased legally. Meanwhile, each free copy can mean a lost sale and lost royalties for the author. While digital rights management locks can hinder simple copying, a determined uploader can remove them, which is why enforcement generally begins after a leak has occurred.
How the Industry Responds: Courts, Registries, and Search
The industry uses three primary tools to combat book piracy, and each has a drawback.
The case progressed rapidly. On May 19, 2026, Judge Jed Rakoff entered default judgment because the operators had not appeared. He awarded the maximum statutory damages of $150,000 for each of the 130 works . That amounts to $19.5 million in total.
The injunction is more significant since it requires over twenty registries, registrars, and hosts — including Cloudflare — to switch off the remaining domains. This is domain blocking at the source. Yet, in August, TorrentFreak reported that the piracy platform was still accessible via .gl, .pk, and .gd addresses managed by registries outside the United States. Legal pressure on providers such as Cloudflare is forcing operators to change their setup.
Search is also one front in the battle against book piracy. By the end of 2025, Google had verified 749 million removals related to addresses on its platform, about 5 percent of all URLs reported since 2012 on copyright grounds. Nevertheless, when searching by name, the main domain still appeared first. Put simply, delisting websites from search engines removes individual book pages without hiding the main entrance.

Book Piracy News: Enforcement Goes Automated as AI Joins In
The key change is that enforcement is now largely automated. By the end of 2025, rights holders were reporting about 10 million new URLs per week for this single platform. Human teams cannot keep up with the volume. A Dutch firm, Linkbusters , files millions of automated takedown requests on behalf of publishers, so many DMCA notices are now generated automatically.
Court orders have followed the pattern of the music labels’ earlier case against the site, which named some of the same intermediaries. Checking whether blocks are working takes real effort, and rights holders and anti-piracy teams often buy proxy access to view regional versions of a pirate site as local visitors do.
The involvement of AI creates pressure. The complaint states that the operators offered high-speed access to over 140 million texts to companies in China, Russia, and other countries, some of which were developing language models. The site also encourages developers to make an “enterprise-level donation”, and a reported crypto price of $200,000 grants full access.
The complaint cites a California court as finding that Meta downloaded the collection while developing its Llama model. The plaintiffs argue that this reduces the market for licensed AI training data.
Note that these are allegations, and a default judgment was entered because the operators did not answer when called upon in court, so the claims were not tested at trial.
Book piracy affects authors most directly. The availability of free copies reduces sales, and pirated material can also end up in AI systems without permission. If you write, publish, or read books, then a few measures can help:
For authors: Check with your publisher to see whether a monitoring service is already filing takedown requests on your books’ behalf. The Google report indicates that over 1,000 authors and publishers have sent notices regarding this platform.
For readers: Avoid pirating books; instead, buy them or borrow them from legitimate bookstores and library apps.
For everyone: Watch the AI licensing debate. It is likely to be the next big fight over books.
The struggle against book piracy is far from over. Although the courts have handed down substantial judgments, U.S. orders may not be enforceable in foreign registries. This gap is likely to lead to calls for stricter domain-blocking rules, particularly for site-blocking. Progress will come in small steps, with each step making it a bit harder to find pirated
Markus lives in San Francisco, California and is the video game and audio expert on Good e-Reader! He has a huge interest in new e-readers and tablets, and gaming.