Fair use, a legal doctrine that allows for a limited usage of copyrighted materials without permission by the original creators for purposes such as teaching and news sharing, is what a lot of generative model creators claim that their training data is covered by.
合理使用是一项法律原则,允许出于教学、新闻传播等目的,在未经原作者许可的情况下有限使用受版权保护的材料。许多生成式模型开发者声称,他们的训练数据正是受到这一原则保护的。
句子以“Fair use”开头,随后插入较长的定义性修饰信息,末尾又出现“what...claim that...is covered by”的嵌套表达和句末介词。虽然主线可以概括为“Fair use is what...”,但核心关系不容易一次定位。