Structuralism? Functionalism? Behaviorism? Or mechanism? Looking for a better approach to AI

Abstract

Purpose

Research of artificial intelligence (AI), has aimed at making machines intelligent via the simulation of natural intelligence, particularly human intelligence. During the past decades, there have been three major approaches aimed at achieving this goal, namely structuralism, functionalism and behaviorism. Unfortunately, they work separately and contradictorily to a large extent. The purpose of this paper is to present a better and more unified approach.

Design/methodology/approach

The paper analyses each of the three major approaches to AI, describing their advantages and disadvantages. There then follows an attempt to explore a new and more reasonable approach to AI. The new approach should be able to solve all the problems that the existing approaches can solve on one hand and can solve the problems that the existing approaches cannot solve on the other hand.

Findings

It was found that the more reasonable and more powerful approach is the one that directly touches the common and core mechanism of intelligence formation. This is due to the fact that the mechanism of intelligence formation is much more essential than other windows of an intelligent system, such as structure, function, or behavior. It was also found that the common and core mechanism of intelligence formation can be implemented through the information‐knowledge‐intelligence transformation. The third finding is that the three existing approaches are special cases of the mechanism approach under different conditions and can thus be harmoniously unified within the frame of the mechanism approach.

Originality/value

The three findings in the paper: the mechanism approach, the implementation of the mechanism approach, and the unification of the existed three major approaches, are important laws never found before in the literature. The breakthrough of the mechanism approach to AI will be of great significance to both theoretical and practical research in AI in the years to come.

Citation

Zhong, Y. (2008), "Structuralism? Functionalism? Behaviorism? Or mechanism? Looking for a better approach to AI", International Journal of Intelligent Computing and Cybernetics, Vol. 1 No. 3, pp. 325-336. https://doi.org/10.1108/17563780810893437