摘要
生成式人工智能的使用者为了获得其生成内容完整的著作权,往往利用生成式人工智能生成内容外观与人类作品外观在辨识上的困难,将其伪装成人类作品以掩盖生成式人工智能的实质性参与。这一行为严重损害了社会公众和自然人创作者的财产性利益和精神性利益,破坏了创作物市场的稳定秩序和信任关系,法律应当加以规制。民法上的诚实信用原则与经济法上的产品信息披露义务分别为生成式人工智能生成内容的标识义务提供了“自律”与“他律”的法律依据,透明化的创作物市场所蕴含的“共同善”理念则促使个人利益服务于社会共善。保证生成式人工智能相关主体履行标识义务的途径包括:在生成阶段,对生成式视觉艺术产品采取标识模式,在设计算法时保证生成内容的可检测性;在传播阶段,应引导相关主体自主披露信息,并建立信用体系以保障义务的履行。
In order to obtain complete copyright,users of generative artificial intelligence often disguise the generated content as human works to cover up the substantial participation of artificial intelligence,which seriously damages the property rights and spiritual interests of the public and natural creators,and undermines the stable order and trust relationship of the creation goods market.This behaviour should be actively dealt with.The principle of good faith in Civil Law and the obligation of product information disclosure in Economic Law respectively provide the legal basis of“self-discipline”and“heteronomy”for the identification obligation of the generated content of generative artificial intelligence.The concept of“common good”contained in the transparent creation market encourages individual interests to serve the common good of society.The approaches to ensure the implementation of the identification obligations of the relevant subjects of generative artificial intelligence include:in the generation stage,adopt the identification mode of generative visual art products,and ensure the detectability of the generated content when designing the algorithm;in the communication stage,relevant subjects should be guided to disclose information independently,and a credit system should be established to ensure the implementation of obligations.
出处
《法商研究》
CSSCI
北大核心
2024年第4期187-200,共14页
Studies in Law and Business
基金
上海政法学院“问渠源学者”计划资助项目
关键词
生成式人工智能
标识义务
诚实信用原则
共同善
generative artificial intelligence
identification obligation
principle of good faith
common good
作者简介
张继红,上海政法学院教授、上海全球安全治理研究院研究员
