歐美研究第五十三卷第二期

322 歐美研究 Does Artificial Intelligence “Have Thoughts”? An Analysis from Husserl’s Phenomenology Ying-Chien Yang Husserl-Archive, University of Cologne E-mail: yepyoung@gmail.com Abstract As the technology on which artificial intelligence is based becomes more sophisticated, animated human-like interactions and superintelligent computer programs will blur the lines between AI and human intelligence. The question arises, do artificial intelligences really “have thoughts” like humans? In this article I will argue, based on Husserl’s phenomenology, that artificial intelligence “can think,” and that the system can indeed respond, but this does not mean that it really “has thoughts.” Necessary conditions of “having thoughts” are perceptual experience and motivational relations. Such perceptual experiences are embodied and open to the intersubjective world. The difference between artificial intelligence and human intelligence does not lie in the higher level of linguistic capability, but in the more fundamental functions of consciousness. Consciousness has its own continuously immanent constitution, which is constantly in changing in processes of self-organization, and of “becoming.” This means that to genuinely “have thoughts” is to have an embodied life. Key Words: consciousness, perception, cognition, artificial intelligence, Husserl’s phenomenology

RkJQdWJsaXNoZXIy ODg3MDU=