镜像自我识别实验——类脑智能自我意识的本质及道德约束  

Mirror Self-Recognition Experiment:The Essence and Moral Constraints of Self-Consciousness in Brain-Like Intelligence

作　　者：宋春艳[1] SONG Chun-yan(Institute of Philosophy,Hunan Academy of Social Sciences,Changsha 410003,China)

机构地区：[1]湖南省社会科学院(省政府发展研究中心)哲学研究所,长沙410003

出　　处：《自然辩证法研究》2023年第4期17-22,共6页Studies in Dialectics of Nature

基　　金：国家社会科学基金一般项目“人机融合智能的道德风险及其协同治理研究”(20BZX029)。

摘　　要：智能机器人随着功能的强大对道德约束的需求更为强烈。自我意识是道德自律形成的基础和前提,智能机器人自我意识的突破将使其道德自律成为可能。心理学中的镜像自我识别实验是验证自我意识的可靠方法,类脑智能研究建立了受大脑启发的机器人身体自我模型,使类脑智能机器人成功挑战了镜像自我识别实验。然而类脑智能机器人的自我意识只是前反思自我意识,并不具有反思能力;仅处于认知的生理和心理层级,尚未达到高阶的文化层级。因此,类脑智能不能实现道德自律,而只能选择伦理设计的关系模式,即在人机融合中采取诉诸他人认同的路径来实现道德约束。With the powerful function of intelligent robots,there is a stronger demand for moral constraints.Self-consciousness is the basis and premise of the formation of moral self-discipline.The breakthrough of intelligent robot's self-consciousness will make its moral self-discipline possible.The mirror self-recognition experiment in psychology is a reliable method to verify self-awareness.The brain-like intelligent research has proposed a brain-inspired robot bodily self model,which makes the brain-like intelligent robot successfully challenged the mirror self-recognition experiment.However,the self-consciousness of brain-like intelligent robot is only pre-reflective self-consciousness,not with the ability to reflect;it is only at the physical and psychological level of cognition,and has not yet reached the higher cultural level.Therefore,brain-like intelligence cannot realize moral self-discipline,but can only choose the relationship mode of ethical design,adopting the path of appealing to others' recognition in human-machine integration.

关 键 词：镜像自我识别 自我意识 类脑智能 道德自律 关系模式 

分 类 号：B82-054[哲学宗教—伦理学] TP242.6[自动化与计算机技术—检测技术与自动化装置]