作为人工智能下一个关口的意识研究——从加扎尼加的意识学说切入

Consciousness Study as the Next Threshold of Artificial Intelligence from Gazzaniga's Theory of Consciousness

随着近年来人工智能(AI)的惊人进步,AI是否具有意识以及如何构建有意识的AI系统越来越受到学界和公众的关注。要回答这类问题,我们仍然需要或者必须从意识理论中去寻找灵感和答案。立足AI意识问题的当前背景,我们考察了加扎尼加在《意识本能:揭秘脑是如何形成心智的》中提出的意识学说。透过脑异常这个窗口,加扎尼加得到了意识现象学的两个重要论断:意识与体验内容不同并且是可以分开的;意识首先是一种感受。为此,他提出了一个关于意识神经机制的模块—层级架构的主张,并用“开水冒泡”隐喻来形象地说明他的意识观。根据这一思路,构建有意识的AI系统需要我们在底层逻辑上辨明“意识”与“智能”的差异和关系,否则,我们就会在AI系统是否有意识以及AI是否会造成对人类的生存性威胁等一系列理论问题上产生困扰。

With the amazing advance in Artificial Intelligence(AI)in recent years,the question of whether AI is conscious and how to build a conscious AI has received increasing attention.To answer such questions,we need or must look to theories of consciousness for inspiration and answers.This paper reviews Gazzaniga's theory of consciousness proposed in his book The Consciousness Instinct:Unraveling the Mystery of How the Brain Makes the Mind,grounding in the context of the problem of AI's consciousness.Through the window of the abnormal brains,which serve as a valuable clue and guide to deduce the neural mechanisms of consciousness,this paper concludes Gazzaniga's two important assertions of such phenomenology:(1)Consciousness and experienced contents are distinct and can be separated;(2)Consciousness is first and foremost a feeling.It is based on these two phenomenological observations that Gazzaniga proposed a modular-layered architecture about the neural mechanism of consciousness,and a metaphor that illustrates the phenomenology of consciousness and the modular-layered architecture vividly,that is,the"boiling bubbles".Accordingly,building conscious AI systems requires to identify"consciousness"and"intelligence"in the underlying logic;otherwise,inanimate silicon-based machines will never be able to mimic our consciousness.