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
李恒威 1阮泽楠1
作者信息
- 1. 浙江大学哲学学院,浙江杭州 310058
- 折叠
摘要
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A1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.
Abstract
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 theo-ry 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 ab-normal 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 ex-perienced 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 a-bout the neural mechanism of consciousness,and a metaphor that illustrates the phenomenology of conscious-ness 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;other-wise,inanimate silicon-based machines will never be able to mimic our consciousness.
关键词
人工智能/加扎尼加/脑异常/模块—层级架构/"开水冒泡"隐喻/意识引用本文复制引用
基金项目
国家社会科学基金重大项目(22&ZD034)
国家社会科学基金(20BZX045)
科技部科技创新2030"脑科学与人类脑研究"重大项目(2021ZD0200409)
出版年
2024
NETL
NSTL
万方数据