确立人工智能法律主体地位的再思考

Rethinking on Establishing the Legal Subject Status of Artificial Intelligence

随着生成式人工智能逐步迈向通用人工智能,其涌现出的独特的思维和理解能力、打通虚拟与现实边界的可交互性以及向人脑规模发展的模型参数量等特征使其在法律层面上应被赋予相应的独立法律主体资格。目前学者们提出的否定说、代理说、限制说等理论无法在逻辑上实现自洽,也未关注到目前人工智能发展的本质特征所应被赋予的法律意义。从理论探索的角度出发,将人工智能行为体运用法律拟制的方式设定为民法上的限制民事行为能力人,运用监护理论中监护人和被监护人之间的权利义务关系来处理人工智能行为体与其他多方主体之间的法律关系,可以从根本上化解新兴权利主体对多法域带来的制度冲击,为我国人工智能产业的发展探索提供前瞻立法制度构建,为全球范围内处理人工智能相关法律问题提供新思路。

As generative artificial intelligence gradually moves towards general artificial intelligence,its unique thinking and un-derstanding ability,the interactivity that opens up the boundary between virtual and reality,and the number of model parameters that develop to the scale of human brain make it be given the corresponding independent legal subject qualification at the legal level.At present,the theories of negation,agency and legal fiction put forward by scholars cannot be logically self-consistent,nor do they see the legal significance that should be given to the essential characteristics of the current development of artificial intelligence.Theoretical and legislative exploration is made from the guardianship theory,the intelligent actor is regarded as a person with limited capacity for civil conduct in the civil law,and the legal relationship between the intelligent actor and other multi-party subjects is handled by the right and obligation relationship between the guardian and the ward,which can fundamen-tally resolve the institutional impact brought by the emerging right subject on the multi-legal domain.To explore a certain forward-looking legislative model for the development of China's artificial intelligence industry.