摘要
经典人工智能是在传统形而上学的隐喻下建构起来的,它在取得了巨大成就以后,很快就步履维艰。事实上,经典人工智能只是模仿了人类智能的“面子”,但并未真正把握到人类智能的“里子”,即“意向性”。换言之,人工智能与人类智能的真正鸿沟是意识的“意向性”,如果没有意向性,机器就不可能表现出真正的智能。现象学家梅洛-庞蒂等的时间性“意识流”和“意向性理论”为人工智能意向性的发显提供了思考空间,彭罗斯的准晶体量子叠加和量子引力技术则对人工智能意向性的现实可能性进行了摹想。基于形而上学的理论困境与经典人工智能的实践困境的亲缘性和同质性,我们可以尝试着将胡塞尔、梅洛-庞蒂和彭罗斯等的意向性理论和技术映射到人工智能研究中来,以推进人工智能的迭代和转型。
AI was constructed under the metaphor of traditional metaphysics,and after achieving great success,it soon encountered difficulties.In fact,Classical AI only imitates the“appearance”of human intelligence,but has not truly grasped the“essence”of human intelligence,that is“intentionality”.In other words,the real gap between AI and human intelligence lies in the“intentionality”of consciousness.Without intentionality,machines cannot exhibit true intelligence.Phenomenologists such as Merleau Ponty’s temporal“stream of consciousness”and“intentionality theory”provide a space for thinking about the emergence of“intentionality”,while Penrose’s quasi crystal quantum superposition and quantum gravity techniques simulate the realistic possibility of“intentionality”.Based on the affinity and homogeneity between the theoretical dilemmas of metaphysics and the practical dilemmas of Classical AI,we can attempt to map the intentionality theories and technologies of Husserl,Merleau Ponty,and Penrose to the research of AI,promoting iteration and transformation of AI.
出处
《河北学刊》
北大核心
2024年第4期85-91,共7页
Hebei Academic Journal
基金
北京市习近平新时代中国特色社会主义思想研究中心北京工业大学研究基地课题
2021年度北京市社科基金规划项目“具身现象学视域下人工智能范式转型研究”(21ZXC008)。
关键词
意向性
“塞尔问题”
意识流
“动觉经验”
量子叠加
intentionality
“Searle Problem”
stream of consciousness
“kinesthetic experience”
quantum superposition
