Abstract: This paper discusses a significant recent shift in the field of philosophy-the transition from conceptual analysis to conceptual engineering. This represents a move from "armchair philosophy" to an approach that incorporates empirical sciences The paper argues that, in the era of artificial intelligence, conceptual engineering has become particularly crucial. Many traditional concepts, such as intelligence, knowledge, consciousness, autonomy, and value, have become ambiguous in both intension and extension. The method of conceptual engineering involves clarifying the reasonable connotations and practical extensions of concepts across various applied domains, using "family resemblance" as an integrative principle to "refine and process" concepts in need of expansion. This results in new representational tools for thinking, reasoning, judging, communicating, and disseminating ideas. In the field of artificial intelligence, conceptual engineering will help clarify the essential distinctions between natural and artificial intelligence, appropriately apply relevant concepts to reflect on and evaluate new phenomena and issues arising from human-machine interaction, and thereby aid in recognizing and leveraging the respective strengths of natural and artificial intelligence.