Discussion of Human Inner Diversity and Typology through Artificial Intelligence Research

Abstract

Recent advances in artificial intelligence research, particularly in generative AI, have greatly expanded our understanding of both AI technology's capabilities and the fundamental nature of 'intelligence'. Traditionally, understanding of 'intelligence' was rooted in the logical validity of algorithmic computation and rule-based symbolic processing. Now, it is increasingly recognized that the essence of intelligence lies in neural network adjustments, which involve pattern extraction from extensive learning data and weighting between patterns, a process not always aligned with logic. Research in large language models further shows that 'intelligence' can intricately learn and utilize concepts without physical experience. This study builds on these AI advancements to challenge the unidimensional perspective of intelligence that emphasizes symbolic methods (abstraction and logic) and explores constructing a scientific model to understand human inner diversity and typology, such as personality and values. The study also considers insights from analytical psychology, assessing the relevance of concepts like the unconscious and intuition in the context of neuroscience and AI research findings.