AI Search represents a vertical application of large language models in the field of information retrieval, providing a novel conversational information acquisition experience which is reshaping human information consumption patterns. This paper delineates the evolutionary trajectory of information acquisition methodologies driven by media technology advancements and identifys three developmental phases: 1) the human-centered information acquisition and memorization, 2) the internet-assisted search paradigm, and 3) the emerging human-AI collaborative interaction empowered by large language models. Two fundamental shifts underpinning this evolution are proposed: first, the transformation of human roles from comprehensive executor to partial executor, and ultimately to collaborator and decision-maker; second, the evolution of human-machine relationships from passive responsiveness to proactive intelligence, culminating in human-AI collaboration. As AI search agents, acting as autonomous agentic entities, increasingly engage in communication and dissemination practices, this paper also critically examines the potential challenges this trend poses to human cognitive capabilities and the broader information ecosystem.