韩国出版商呼吁加强版权保护和赔偿，因疑似与人工智能训练有关的大批量购书请求引发了对知识产权盗窃的担忧。
South Korean publishers are calling for stronger copyright protections and compensation as bulk book purchase requests suspected of being linked to artificial intelligence training raise concerns over intellectual property theft.
出版团体正在调查可疑订单并寻求协同应对。然而，作者和出版商面临一个根本性障碍：他们往往无法确切知道自己的书籍是否被用于训练AI系统。
Publishing groups are investigating suspicious orders and seeking a coordinated response. However, authors and publishers face a fundamental obstacle: they often cannot know for sure whether their books are being used to train AI systems.
韩国连锁书店教保文库周三表示，自今年早些时候起，包括美国和英国企业在内的10家海外公司曾接洽该公司，每家仅寻求购买约10万种书名的单本。
Kyobo Book Centre, a bookstore chain in South Korea, said on Wednesday that 10 overseas companies, including businesses in the United States and Britain, had approached it since early this year, each seeking only one copy of roughly 100,000 titles.
据韩联社援引教保文库一名代表的话报道：“许多公司自称是图书批发商。”
“Many introduced themselves as book wholesalers,” a Kyobo representative was quoted as saying by Yonhap News Agency.
该代表补充称，部分公司否认是为了AI训练目的，但指出最终接收方是数据相关公司。因无法排除用于AI训练的可能性，该书商拒绝了该请求。
The representative added that some denied an AI training purpose but identified data-related companies as the final recipients. The bookseller declined the request because it could not rule out the use of AI training.
另一家书店Yes24据报也在8月下旬收到一家美国分销商关于大批量购买二手书的询价。双方举行过一次会面，但讨论在该分销商停止回应后中止。
Yes24, another bookstore, also reportedly received an inquiry from a US distributor in late August about buying used books in bulk. The two sides held one meeting, but discussions stopped after the distributor ceased responding.
上周，韩国文学翻译院表示，在确认21部本地作者作品出现在一份与美国版权和解案涉及的Anthropic公司赔偿资格名单上后，已通知相关作者。
Last week, the Literature Translation Institute of South Korea said it notified local authors after identifying 21 of their books on a compensation eligibility list associated with a US copyright settlement involving Anthropic.
这家AI公司面临作者发起的集体诉讼，指控其非法下载受版权保护的作品用于模型训练。
The AI company faces a class-action lawsuit by authors over allegations that it illegally downloaded copyrighted works for model training.
在日本，哲学、历史和医学类二手书的大批量采购也引发了怀疑，认为是AI开发商在获取材料用于扫描。
In Japan, bulk purchases of second-hand books on philosophy, history and medicine have also fuelled suspicions of AI developers acquiring material for scanning.
本周，韩国出版协会已开始对书店、发行商、旧书商和出版社进行调查，了解可能与人工智能相关的大宗订单情况。一位协会官员表示，调查结束后，出版团体应组建联合机构，旨在建立补偿安排和立法。
This week, the Korean Publishers Association has begun surveying bookstores, distributors, second-hand booksellers and publishers about potentially AI-related bulk orders. An association official said publishing organisations should form a joint body after the survey that will aim at establishing compensation arrangements and legislation.
与此同时，韩国出版学会发出呼吁，要求在出版物被用于人工智能训练时给予公平补偿。截至周三，已有240多家出版社和团体，以及702名作家、译者、插画家和出版从业者加入该呼吁。
The Korean Publishers Society, meanwhile, issued an appeal demanding fair compensation when publications are used for AI training. Over 240 publishers and organisations, along with 702 authors, translators and illustrators and publishing professionals, joined the appeal as of Wednesday.
出版社Prunsoop的代表赵韩娜（Cho Han-na，音）表示，单个权利人无法单独有效地对抗大型海外科技公司。
Cho Han-na, head of publisher Prunsoop, said individual rights holders could not effectively confront large overseas technology companies alone.
出版商还主张，他们缺乏关于人工智能公司从何处获取书籍内容以及如何使用这些内容的信息。
Publishers also argue they lack information about where AI companies obtain book content and how they use it.
德成女子大学AI DynaInfo研究院的版权专家、教授朴贞仁（Park Jung-in，音）表示，信息差距使得协商许可条款或证明涉嫌侵权变得困难。
Park Jung-in, a copyright specialist and professor at Duksung Women’s University’s AI DynaInfo institute, said information gaps made it difficult to negotiate licensing terms or prove suspected infringement.
朴贞仁说：“核心问题是一种权利虽存在但在实践中难以行使的结构，”她呼吁建立程序，让权利人能够核查训练数据的来源和使用情况。据《韩国时报》报道。
“The core problem is a structure in which rights exist but are difficult to exercise in practice,” Park said, calling for procedures that would allow rights holders to check the sources and use of training data. By The Korea Times