人工智能刑法讨论的有效性探析:拆解、批判与步骤

Analysis of the Effectiveness of Artificial Intelligence Criminal Law--Divide,Criticize and Step

人工智能刑法呈现出多维特征,存在现实价值,但是目前的讨论缺少统一范式和上层反思。“人工智能刑法讨论何以有效”的命题得以提出,并结合目前主流的研究范式,可以拆解为:一是场景讨论可否独立支撑起人工智能刑法讨论的有效性;二是场景讨论是否触及体系讨论;三是如何展开人工智能刑法的讨论。第一个问题的答案是否定的,自动驾驶汽车、游戏外挂以及个人信息等成熟场景都进一步引发了更多值得讨论的问题,触及了第二个问题的回答,即体系化的必要性。通过场景讨论部分的反思,我们可以得出三步人工智能刑法的未来发展方向:细化场景讨论,进一步分析原理沟通的可能;以可解释性为中心,双向促进人工智能刑法的发展;全时段、全过程促进人工智能刑法共同体的建立。

Artificial intelligence criminal law embodies multi-dimensional characteristics and realistic value,but the current discussion about it lacks a unified paradigm and upper level reflection.Therefore,the proposition of"How Is the AI Criminal Law Discussion to be Effective"is put forward.And combined with the current mainstream research paradigm,it can be divided into three parts:first,whether the scenario discussion can independently support the effectiveness of the discussion of artificial intelligence criminal law;second,whether the scenario discussion connects to the system discussion;third,how to carry out the discussion of artificial intelligence criminal law.The answer to the first question is no.Mature scenarios such as autonomous vehicle,game plug-ins and personal information further raise more questions worthy of discussion and connect to the answer to the second question,that is,the necessity of systematization.Through reflection on the scenario discussion part,we can draw three steps towards the future development direction of artificial intelligence criminal law:Refine the scenario discussion to further analyze the possibility of principle communication;Bi-directionally promote the development of artificial intelligence criminal law with interpretability as the center;Promote the establishment of artificial intelligence criminal law community in the whole period and process.