Abstract: Generative artificial intelligence (AI) exhibits a level of intelligence that traditional machines do not possess. Does this mean that the labor theory of value is no longer applicable? Generative AI is essentially a probabilistic computational model: it lacks the conscious purposiveness unique to human labor and does not possess unlimited adaptability, and it is the result of a further separation between the conception and execution of human labor. In the real economy, it primarily substitutes for conventional cognitive work, and its application scenarios are confined to industries with a high degree of standardization and relatively fixed workflows. Generative AI activities lack social subjectivity, are not part of abstract social labor, and do not create new value. Generative AI can only gradually transfer the “old value” embedded in itself into new products. The application of generative AI indeed raises individual labor productivity, and departments that adopt it capture value transferred from other departments; these values collectively constitute the source of value for generative AI products.

Key words: generative artificial intelligence, labor theory of value, value creation, value transfer

摘要： 生成式人工智能展现出了传统机器所不具有的智能水平，这是否意味着劳动价值论已经不再适用。生成式人工智能的本质是概率计算模型，缺乏人类劳动所特有的自觉目的性，也不具备无限的适应性，是人类劳动的概念与执行进一步分离的结果。在现实经济中，它主要替代常规型认知工作，其应用场景局限于标准化程度高、作业流程相对固定的行业。生成式人工智能的活动不具有社会主体性，不是社会抽象劳动的一部分，不创造新价值。生成式人工智能只能将自身所包含的“旧价值”逐步转移到新产品中。生成式人工智能的应用确实会提高个别劳动生产率，使用生成式人工智能的部门获得来自其他部门的价值转移，这些价值共同构成了生成式人工智能产品的价值来源。

关键词: 生成式人工智能, 劳动价值论, 价值创造, 价值转移