The pursuit of Artificial General Intelligence (AGI) represents one of the most ambitious goals in artificialintelligence research, yet current approaches often overlook the foundational principles that govern biological intelligence.This paper introduces and examines Machine Psychology as an interdisciplinary framework that systematically integratesprinciples from learning psychology-particularly operant conditioning and Relational Frame Theory-with adaptive reasoningsystems to advance AGI development. We propose a bidirectional learning model wherein psychological principlesinform AI architecture while AI systems provide novel insights into cognitive mechanisms. Through analysis of recentimplementations using the Non-Axiomatic Reasoning System (NARS), we demonstrate how core psychological constructssuch as reinforcement learning, derived relational responding, and functional equivalence can be computationally realizedto produce flexible, context-sensitive artificial cognition. This framework addresses critical limitations in contemporaryAI systems, including brittleness in novel contexts, inability to generalize across domains, and lack of metacognitivecapabilities. The paper further explores implementation challenges specific to developing economies, using Uzbekistan’saccelerated digital transformation as a case study for culturally-adapted AGI development strategies. We conclude thatMachine Psychology offers a principled pathway toward human-level artificial intelligence while simultaneously enrichingour understanding of natural cognition.