This chapter brings conclusions to the findings of the previous chapters and outlooks on their implications for the orientation of AI development. A computational functionalist account of thinking, using data structures plus algorithms to model mental representations plus computational procedures, can effectively reveal the functioning of thinking, but fails to explain how data structures and computational procedures are formed. By demonstrating how the internet and semantic technologies enhance both human and artificial information processing agents, the conditions of possibility and normative imperatives for rational action by information processing agents are revealed from an inferentialist perspective. This reveals the book’s central claim that the very idea of AI needs to be modeled within a normative framework, in which the rational agency of interacting between agents and their shared environments forms and regulates data structures and information processing procedures. The explanation of intelligent agents should accordingly move from the uncovering of their internal states to the examination of their normative statuses as they engage in digital social practices. This chapter will expand on the significance of this shift in three aspects for understanding and developing AI in digital social practices.