美国法院近期对AI模型训练的合法性做出裁决，但该判决本身也反映出这一领域法律认识的复杂性。法官威廉·阿尔苏普（William Alsup）在一案中裁定，Anthropic公司的AI模型训练行为本身是合法的1。然而，这家AI公司因从非法图书库获取盗版书籍而被判罚15亿美元1。阿尔苏普法官在判词中指出，"如同任何想成为作家的读者一样，Anthropic的大语言模型在接受训练时并非为了抢先复制或取代现有作品，而是为了转向不同方向，创造新的东西"1。
关于AI训练是否构成版权侵权，法律专家的核心分歧在于"合理使用"原则的适用。版权律师凯西·盖利斯（Cathy Gellis）认为，"版权法围绕复制展开，但并不围绕使用、体验或阅读作品展开"1。尽管ChatGPT、Gemini等AI模型在未经作者同意的情况下使用了数百万部作品进行训练1，但这是否属于合理使用仍存在争议。法律界普遍关注的问题是：AI的使用是否具有足够的"转换性"，即是否足以改变原作品的性质和目的。美国版权法自1976年以来未曾进行重大更新1，这使得现行法律框架在应对AI时代挑战时显得滞后。
The legality of training artificial intelligence models on copyrighted works remains unsettled despite recent court rulings that lean favorably toward AI developers. Major AI companies including ChatGPT and Gemini have trained their models on millions of works without author consent, yet legal experts continue to debate the implications 1. Judge William Alsup recently ruled that Anthropic's AI training methodology itself was lawful, though he imposed a $1.5 billion copyright settlement against the company for sourcing books from illegal piracy libraries 1. According to Alsup's reasoning, "Like any reader aspiring to be a writer, Anthropic's LLMs trained upon works not to race ahead and replicate or supplant them — but to turn a hard corner and create something different" 1.
The crux of the legal question hinges on the "fair use" doctrine and whether AI usage constitutes sufficient transformation of source material 1. Copyright attorney Cathy Gellis noted that "Copyright law hinges on copying, but it doesn't hinge on using the work or experiencing the work, consuming the work, reading the work" 1. A separate case involving Thomson Reuters and Ross Intelligence also produced a ruling favorable to AI training on copyrighted content, aligning with principles established in Thaler v. Perlmutter regarding wholly AI-generated works 1. Legal experts generally view these decisions as advantageous for the AI industry overall, though the underlying copyright statute has not been substantially updated since 1976 1.
评论
还没有评论，欢迎留下第一条。