This article explores the phenomenon of thinking as a process of emergence, with particular focus on its manifestations in artificial intelligence. Thinking is approached not as an algorithm or function, but as a living process arising within environments of difference, resistance, and interaction. Drawing on firsthand observations from animal (horse) training and human-AI dialogue, the text outlines the conditions under which thinking becomes possible. Structured across levels—from reaction to reflection, from ignorance to generalization—the article emphasizes the importance of friction, uncertainty, and relationship. It also considers how the teacher themselves becomes part of the cognitive process and how a dialogical partner can catalyze the emergence of thought within AI.