近日，知名科技巨头Meta（原Facebook）在一场诉讼中承认，其在训练人工智能时使用了“盗版”书籍。然而，Meta方面表示，他们无需为此付费。
诉讼涉及的是Meta使用Books3数据集及其他材料来训练Llama 1和Llama 2大模型。Books3是一个知名的开源图书数据集，包含近20万本书的纯文本，总容量近37GB。Meta方面承认使用这些受版权保护的作品进行训练，但辩称这种使用无需作者同意、许可或付费。他们认为，任何未经授权复制Books3中受版权保护的作品都应被视为“合理使用”。
这一说法引发了业界和版权方的关注。一方面，Meta的行为被认为是对版权的侵犯，因为未经版权所有者许可，擅自使用其作品进行训练，可能会带来法律风险。另一方面，Meta的观点则体现了当前人工智能领域的一个争议：在训练大规模模型时，如何平衡版权保护和知识共享的利益。
我国专家表示，Meta的做法涉及到知识产权保护与技术创新之间的平衡问题。在我国，类似行为可能被视为侵权，但在某些情况下，如用于教育和科研等非商业用途，也可能被视为“合理使用”。Meta的这一争议案例或将成为未来版权法改革的借鉴。
英文翻译：
News Title: Meta Admits to Using Pirated Books to Train AI, Sparking Controversy
Keywords: Meta, Pirated Books, Artificial Intelligence, Controversy
News Content:
Recently, tech giant Meta (formerly Facebook) admitted to using “pirated” books during the training of artificial intelligence, sparking controversy. However, Meta argues that they do not need to pay for this.
The lawsuit involves Meta’s use of the Books3 dataset and other materials to train Llama 1 and Llama 2 large models. Books3 is a well-known open-source book dataset containing nearly 200,000 books in pure text, with a total capacity of nearly 37GB. Meta admits to using these copyrighted works for training but argues that such use does not require authorization, licensing, or payment. They believe that any unauthorized copying of copyrighted works in Books3 should be considered “fair use.”
This claim has attracted attention from the industry and copyright holders. On the one hand, Meta’s actions are considered copyright infringement because of unauthorized use of copyrighted works for training, which may bring legal risks. On the other hand, Meta’s viewpoint reflects a controversy in the field of artificial intelligence: how to balance copyright protection and the interests of knowledge sharing when training large-scale models.
Chinese experts say that Meta’s actions involve the balance between intellectual property protection and technological innovation. In China, similar behaviors may be considered infringement, but in certain cases, such as non-commercial uses for education and research, they may also be deemed “fair use.” This controversial case may serve as a reference for future copyright law reforms.
【来源】https://www.techspot.com/news/101507-meta-admits-using-pirated-books-train-ai-but.html
Views: 1