未来人工智能是否能够跨越“恐怖谷”？——一种现象学视角的考察

Will AI be able to cross “the uncanny valley”?——An investigation from the perspective of phenomenology

智能机器人能模仿人到什么程度,或者说智能机器人的逼真度究竟能够达到什么高度?这很大程度上可以看作是对未来人工智能的极限问题的另一种提法。智能机器人的逼真度会引出智能机器人相关的一个基本理论,即恐怖谷理论。通过现象学的理论视角,结合人工智能哲学的基本理论,以恐怖谷理论为切入点,探讨智能机器人的逼真度、人工智能的学科基础和未来的可能限度等问题。

To what extent can intelligent robots imitate people,or what level can the fidelity of intelligent robots achieve?To a great extent,it can be regarded as another formulation of the limit problem of artificial intelligence in the future.The fidelity of intelligent robot will lead to a basic theory related to intelligent robot,that is,the theory of uncanny valley.In this paper,from the perspective of phenomenology and the basic theory of artificial intelligence philosophy,starting from the analysis of the uncanny valley,discusses the fidelity of intelligent robot,the discipline basis of artificial intelligence and the possible limits in the future.