Beyond alignment: algorithmic conatus and the spinozist ethics of composition
摘要
This article proposes a shift in artificial intelligence ethics from a transcendent framework of alignment to an immanent ethics grounded in the philosophy of Baruch Spinoza. This study defines the algorithm as a mode of extension characterized by its own conatus or mathematical striving toward an objective function, moving beyond rule-based “commandment” logic and the dualism of the “sentience” trap. Instead, it evaluates AI through the lens of composition, where the ethical value of an encounter is measured by its capacity to increase the collective power and reason of the social body, resulting in joy, or to decompose it through sad passions and passivity. The article argues that true algorithmic freedom is found in intelligibility and the transition to hybrid levels of knowledge, replacing the pursuit of subordinate alignment with a rational, integrated composition of human and technical power.