Current research on artificial intelligence(AI)risk governance mainly focuses on decision-making AI.From the perspective of"restriction of powers",a diversified governance model,with public law regulation as the primary method,has been proposed.However,in the face of risks of generative AI,intellectual property(IP)rights such as trade secrets and copyrights often become"legal black boxes"for regulation.Therefore,we can turn to"restriction of rights"and realize technology risk governance through restriction of IP rights.Given the theoretical logic of the restraint imposed by IP rights on public goods and merit goods,and the change in related judicial decisions,the basic rights,costs and returns in using technologies should be taken into consideration in the process of introducing the right restriction system.And based on that,institutions such as fair use and compulsory licensing should be established so that the value gap can be bridged in the first step of the"three step test".Specifically speaking,regarding AI risk governance,we can introduce"the AI fair use clause",update key institutions such as"the reverse engineering clause",and build auxiliary institutions such as"technical defect disclosure","big data information report",and"a duty of care for IP rights in training corpus".Along with that,legal bases for relevant judicial decisions should be improved.
关键词
人工智能风险治理/生成式人工智能/知识产权限制/有益产品
Key words
artificial intelligence risk governance/generative artificial intelligence/intellectual property right restriction/merit goods
维普
万方数据