Why Are We Hackable? Three Channels of Infiltration in Learning Systems: From Viruses to Algorithms

This article investigates the fundamental question of why the vertebrate brain (specifically the human brain) possesses inherent vulnerabilities that allow for "possession" by both biological and informational agents. Building on the concept of "functional convergence," the author argues that susceptibility to hacking is an unavoidable byproduct of an adaptive evolutionary architecture designed for social learning and environmental flexibility. The paper introduces a unified framework comprising three specific "neural channels of infiltration": the manipulation of prediction error, the exploitation of prestige bias, and the hijacking of the reward system. By synthesizing evidence from biological parasitology, memetic theory, and Large Language Models (LLMs), this study offers a cross-disciplinary perspective on addiction, ideological radicalization, and AI safety.