强人工智能刑事责任主体地位之再肯定

Reaffirmation on Artificial General Intelligence's Status as Subject of Criminal Responsibility

学习算法及深度学习理论说明强人工智能具有存在的可能性。与弱人工智能不同,强人工智能行为的不可预测性、脱离人类主体控制的特点,决定必须将强人工智能作为独立法律主体加以规制,这符合理性人假设,说明有法律规制的必要性和可行性。基于对现行刑法刑事责任主体资格和共同犯罪理论的分析,结合强人工智能成长发展的变化规律,将强人工智能分为无刑事责任、限制刑事责任和完全刑事责任,强人工智能实施犯罪行为完全符合现行刑法规定,应肯定强人工智能的刑事责任主体地位。

Learning algorithms and the Deep Learning Theory have shown that Artificial General Intelligence(AGI)has possibility of existence.Different from Artificial Narrow Intelligence(ANI),AGI has unpredictable behavior and the feature to escape from the control of human beings.Therefore,AGI shall be regulated as an independent legal subject,which not only conforms to the assumption of a rational person,but has the regulatory necessity and feasibility.Based on analyzing the status as subject of criminal responsibility and the common crime theories in the current criminal law,as well as the change rules in the growth and development of AGI,this article classifies AGI into types which shall bear no criminal responsibility,limited criminal responsibility or complete criminal responsibility,and then points out that the criminal behaviors of AGI completely comply with the provisions of current criminal law,so its status as subject of criminal responsibility shall be affirmed.