生成式人工智能数据来源的合法性探究

中图分类号：D913 文献标志码：A 文章编号：2097-3853(2025)05-0482-06

Research on legitimacy of data sources of generative artificial intelligence

LIU Xia，FANG Baolun (School ofLaw and Intellectual Property，Fujian Universityof Technology，Fuzhou 350118,China)

Abstract: In recent years，the generative artificial inteligence technology represented by ChatGPT has developed rapidly.At thesame time，its data source also implies legal risks such as infringement of personal information and intellectual property rights and improperacquisition methods.Against the background of existing relevant infringement cases，according to the actual situation of data source infringement of generative artificial intellgence，personal information protection system and intellectual property laws and regulations can be applied，and relevant legal normsand principles can be embedded in the fieldof data source of generative artificial intelligence to regulate theacquisition anduseof data source of generative artificial intelligence. On this basis，the responsibilitiesof relevant subjects can be clarified，and the technologies of data collction, processing anduse can be improved，the legitimacy of artificial inteligence data sources can be standardized byconstructing corresponding supervision modes within enterprises， between industries and government departments，so as to better solve the challenges brought by new artificial intelligence technologies.

Keywords : generative artificial inteligence；data source；reasonable use；safe harbor rules；whole-chain supervision mode

随着人工智能技术的迅速发展，数据作为一种重要资源对生成式人工智能（AIGC）训练模型构建和深度学习起着至关重要的作用。（剩余9559字）