本文為回應 AI 擴張下的「消極不自由」，從自由主義與科學哲學立場，檢視演算法宰制與資本對個體行動邊界的壓縮，並提出以宋明理學「定靜」工夫作為安頓之道。首先，本文以爲AI 造成的不自由樣態，指出其中的推薦機制，將造就欲望的重塑，並強化工具的依賴，可能引發社會與政治的各種風險。其次，針對不自由的狀態，本文嘗試從理學「定靜」工夫裡的「內靜外敬」思維，轉譯為應對AI 工具的原則。「內靜」是「心」的不拘執於「物」，「外敬」則是「心」對「物」的自持約束，藉此設定與AI工具的距離與邊界，以還原「物」的工具性。並進一步強調「定靜」工夫的新義，將不以古典哲學的「內聖」作為導向，而將以重獲「消極自由」為目的。經由整合政治哲學、科學哲學與理學工夫論，進行跨域探究，從中提出應對時代問題工夫的可能框架。由是言之，本文雖是初步、但卻有一定新意的嘗試。

This paper responds to the condition of “negative unfreedom” under the expansion of AI. From the perspectives of liberalism and the philosophy of science, it examines how algorithmic domination and capital compress the boundaries of individual action, and it proposes the Song–Ming Neo-Confucian practice of “dingjing” (定靜) as a mode of anchoring. First, the paper argues that AI-induced unfreedom emerges through recommender mechanisms that reshape desire and deepen dependence on tools, thereby generating a range of social and political risks. Second, addressing this unfreedom, it translates the “inner stillness and outer reverence” (內靜外敬) framework of dingjing into principles for engaging AI tools: “inner stillness” denotes the mind’s nonattachment to things, while “outer reverence” denotes the mind’s selfdisciplined restraint toward things. On this basis, the user’s distance from and boundaries with AI tools can be set so as to restore the instrumentality of things. The paper further advances a renewed interpretation of dingjing, oriented not toward the classical ideal of “inner sageliness” (內聖) but toward the recovery of negative freedom. By integrating political philosophy, philosophy of science, and Neo-Confucian discourse on self-cultivation, this cross-disciplinary inquiry proposes a tentative yet meaningful practical framework for addressing contemporary problems.