Analogy-Made Humanized AI
Mark Chang · 2025
This chapter delves into the architecture and principles of humanized AI (HAI), highlighting its capacity to replicate human cognition, learning, and behavior through adaptive and dynamic mechanisms. Unlike mainstream AI approaches that rely heavily on big data, HAI adopts a growing-data approach, inspired by the incremental learning process of a human child. This framework integrates key principles such as attention mechanisms, hierarchical tokenization, recursive patternization, and similarity-based learning , enabling HAI agents to adapt and generalize effectively, albeit at a slower pace, with the potential to learn anything in principle. By incorporating innate knowledge, sensory embodiment, and self-awareness , HAI systems develop robust and scalable learning mechanisms that prioritize flexible and meaningful interactions.